Merge pull request 'FC-3h + FC-3k: backup first-class + admin destructive actions' (#10) from dev into main

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

alembic/versions/0016_fc3i_task_run.py

@@ -0,0 +1,86 @@
+"""fc3i: task_run table
+
+Revision ID: 0016
+Revises: 0015
+Create Date: 2026-05-24
+
+Additive only. New table records every Celery task attempt via signal
+handlers (backend.app.celery_signals). Status is plain String(16) not
+Postgres ENUM (per feedback_check_existing_enums: ENUM columns hard-
+fail at INSERT, String columns extend cleanly).
+
+Composite indexes anticipate the three dashboard panes:
+- (queue, started_at desc)   — per-lane recent activity
+- (status, started_at desc)  — recent failures pane
+- (task_name, started_at desc) — drill-down by task
+
+Indexed columns get individual indexes via `index=True` on the model;
+the composites below cover the multi-column lookups.
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0016"
+down_revision: Union[str, None] = "0015"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "task_run",
+        sa.Column("id", sa.Integer(), primary_key=True),
+        sa.Column("celery_task_id", sa.String(length=64), nullable=False),
+        sa.Column("queue", sa.String(length=32), nullable=False),
+        sa.Column("task_name", sa.String(length=128), nullable=False),
+        sa.Column("target_id", sa.Integer(), nullable=True),
+        sa.Column("started_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("finished_at", sa.DateTime(timezone=True), nullable=True),
+        sa.Column("duration_ms", sa.Integer(), nullable=True),
+        sa.Column(
+            "status", sa.String(length=16), nullable=False,
+            server_default="running",
+        ),
+        sa.Column("error_type", sa.String(length=128), nullable=True),
+        sa.Column("error_message", sa.Text(), nullable=True),
+        sa.Column("retry_count", sa.Integer(), nullable=True),
+        sa.Column("worker_hostname", sa.String(length=128), nullable=True),
+        sa.Column("args_summary", sa.String(length=255), nullable=True),
+    )
+
+    # Single-column indexes (matches Mapped[...].index=True on model).
+    op.create_index("ix_task_run_celery_task_id", "task_run", ["celery_task_id"])
+    op.create_index("ix_task_run_queue", "task_run", ["queue"])
+    op.create_index("ix_task_run_task_name", "task_run", ["task_name"])
+    op.create_index("ix_task_run_started_at", "task_run", ["started_at"])
+    op.create_index("ix_task_run_finished_at", "task_run", ["finished_at"])
+    op.create_index("ix_task_run_status", "task_run", ["status"])
+
+    # Composite indexes for dashboard query patterns.
+    op.create_index(
+        "ix_task_run_queue_started",
+        "task_run", ["queue", sa.text("started_at DESC")],
+    )
+    op.create_index(
+        "ix_task_run_status_started",
+        "task_run", ["status", sa.text("started_at DESC")],
+    )
+    op.create_index(
+        "ix_task_run_name_started",
+        "task_run", ["task_name", sa.text("started_at DESC")],
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_task_run_name_started", table_name="task_run")
+    op.drop_index("ix_task_run_status_started", table_name="task_run")
+    op.drop_index("ix_task_run_queue_started", table_name="task_run")
+    op.drop_index("ix_task_run_status", table_name="task_run")
+    op.drop_index("ix_task_run_finished_at", table_name="task_run")
+    op.drop_index("ix_task_run_started_at", table_name="task_run")
+    op.drop_index("ix_task_run_task_name", table_name="task_run")
+    op.drop_index("ix_task_run_queue", table_name="task_run")
+    op.drop_index("ix_task_run_celery_task_id", table_name="task_run")
+    op.drop_table("task_run")

alembic/versions/0017_fc3h_backup_run.py

@@ -0,0 +1,82 @@
+"""fc3h: backup_run table
+
+Revision ID: 0017
+Revises: 0016
+Create Date: 2026-05-24
+
+Additive. New table records every backup/restore attempt with artifact
+metadata. Lifecycle tracking lives in task_run from FC-3i; this is
+artifact-only (paths, sizes, tag, restore lineage).
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0017"
+down_revision: Union[str, None] = "0016"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "backup_run",
+        sa.Column("id", sa.Integer(), primary_key=True),
+        sa.Column("kind", sa.String(length=16), nullable=False),
+        sa.Column(
+            "status", sa.String(length=16), nullable=False,
+            server_default="pending",
+        ),
+        sa.Column("tag", sa.String(length=64), nullable=True),
+        sa.Column("triggered_by", sa.String(length=32), nullable=False),
+        sa.Column("started_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("finished_at", sa.DateTime(timezone=True), nullable=True),
+        sa.Column("sql_path", sa.Text(), nullable=True),
+        sa.Column("tar_path", sa.Text(), nullable=True),
+        sa.Column("size_bytes", sa.BigInteger(), nullable=True),
+        sa.Column("error", sa.Text(), nullable=True),
+        sa.Column(
+            "manifest", sa.JSON(), nullable=False, server_default="{}",
+        ),
+        sa.Column(
+            "restored_from_id", sa.Integer(),
+            sa.ForeignKey("backup_run.id", ondelete="SET NULL"),
+            nullable=True,
+        ),
+    )
+
+    # Single-column indexes (matches Mapped[...].index=True).
+    op.create_index("ix_backup_run_kind", "backup_run", ["kind"])
+    op.create_index("ix_backup_run_status", "backup_run", ["status"])
+    op.create_index("ix_backup_run_tag", "backup_run", ["tag"])
+    op.create_index("ix_backup_run_started_at", "backup_run", ["started_at"])
+    op.create_index("ix_backup_run_finished_at", "backup_run", ["finished_at"])
+
+    # Composite indexes for dashboard query patterns.
+    op.create_index(
+        "ix_backup_run_kind_started",
+        "backup_run", ["kind", sa.text("started_at DESC")],
+    )
+    op.create_index(
+        "ix_backup_run_status_finished",
+        "backup_run", ["status", sa.text("finished_at DESC")],
+    )
+    # Partial index: only tagged rows participate in retention-exempt query.
+    op.create_index(
+        "ix_backup_run_tag_partial",
+        "backup_run", ["tag"],
+        postgresql_where=sa.text("tag IS NOT NULL"),
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_backup_run_tag_partial", table_name="backup_run")
+    op.drop_index("ix_backup_run_status_finished", table_name="backup_run")
+    op.drop_index("ix_backup_run_kind_started", table_name="backup_run")
+    op.drop_index("ix_backup_run_finished_at", table_name="backup_run")
+    op.drop_index("ix_backup_run_started_at", table_name="backup_run")
+    op.drop_index("ix_backup_run_tag", table_name="backup_run")
+    op.drop_index("ix_backup_run_status", table_name="backup_run")
+    op.drop_index("ix_backup_run_kind", table_name="backup_run")
+    op.drop_table("backup_run")

alembic/versions/0018_fc3h_backup_settings.py

@@ -0,0 +1,62 @@
+"""fc3h: backup_* knobs on import_settings
+
+Revision ID: 0018
+Revises: 0017
+Create Date: 2026-05-24
+
+Adds four columns to the singleton import_settings row:
+  - backup_db_nightly_enabled (default False — opt-in)
+  - backup_db_nightly_hour_utc (default 3)
+  - backup_db_keep_last_n (default 14)
+  - backup_images_keep_last_n (default 3)
+
+server_default ensures the singleton row is backfilled in place
+without an UPDATE statement.
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0018"
+down_revision: Union[str, None] = "0017"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "import_settings",
+        sa.Column(
+            "backup_db_nightly_enabled", sa.Boolean(),
+            nullable=False, server_default=sa.false(),
+        ),
+    )
+    op.add_column(
+        "import_settings",
+        sa.Column(
+            "backup_db_nightly_hour_utc", sa.Integer(),
+            nullable=False, server_default="3",
+        ),
+    )
+    op.add_column(
+        "import_settings",
+        sa.Column(
+            "backup_db_keep_last_n", sa.Integer(),
+            nullable=False, server_default="14",
+        ),
+    )
+    op.add_column(
+        "import_settings",
+        sa.Column(
+            "backup_images_keep_last_n", sa.Integer(),
+            nullable=False, server_default="3",
+        ),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("import_settings", "backup_images_keep_last_n")
+    op.drop_column("import_settings", "backup_db_keep_last_n")
+    op.drop_column("import_settings", "backup_db_nightly_hour_utc")
+    op.drop_column("import_settings", "backup_db_nightly_enabled")

backend/app/api/__init__.py

@@ -14,6 +14,7 @@ api_bp.add_url_rule("/health", view_func=health.get_health, methods=["GET"])
 
 
 def all_blueprints() -> list[Blueprint]:
+    from .admin import admin_bp
     from .aliases import aliases_bp
     from .allowlist import allowlist_bp
     from .artist import artist_bp
@@ -33,6 +34,8 @@ def all_blueprints() -> list[Blueprint]:
     from .showcase import showcase_bp
     from .sources import sources_bp
     from .suggestions import suggestions_bp
+    from .system_activity import system_activity_bp
+    from .system_backup import system_backup_bp
     from .tags import tags_bp
     return [
         api_bp,
@@ -44,6 +47,9 @@ def all_blueprints() -> list[Blueprint]:
         artists_bp,
         showcase_bp,
         settings_bp,
+        system_activity_bp,
+        system_backup_bp,
+        admin_bp,
         import_admin_bp,
         migrate_bp,
         suggestions_bp,

backend/app/api/admin.py

@@ -0,0 +1,200 @@
+"""FC-3k: /api/admin — destructive admin actions.
+
+Five action surfaces:
+  POST   /api/admin/artists/<slug>/cascade-delete  (Tier C)
+  POST   /api/admin/images/bulk-delete             (Tier C)
+  DELETE /api/admin/tags/<int:tag_id>              (Tier B)
+  POST   /api/admin/tags/<int:dest_id>/merge       (Tier B)
+  POST   /api/admin/tags/prune-unused              (Tier A)
+  GET    /api/admin/tags/<int:tag_id>/usage-count  (helper)
+
+Tier-C ops take a dry_run body flag (returns projection inline,
+no dispatch) and a confirm body field (server-recomputed token).
+Long-running ops dispatch a maintenance-queue Celery task; the UI
+tails FC-3i's /api/system/activity/runs to surface progress.
+"""
+from __future__ import annotations
+
+import hashlib
+
+from quart import Blueprint, jsonify, request
+from sqlalchemy import select
+
+from ..extensions import get_session
+from ..models import Artist
+from ..services.cleanup_service import project_artist_cascade, project_bulk_image_delete
+
+admin_bp = Blueprint("admin", __name__, url_prefix="/api/admin")
+
+
+def _bad(error: str, *, status: int = 400, **extra):
+    body = {"error": error}
+    body.update(extra)
+    return jsonify(body), status
+
+
+def _bulk_image_confirm_token(image_ids: list[int]) -> str:
+    """Stable 8-hex token derived from the sorted id list. Mutates
+    when the selection changes; stays the same across modal opens of
+    the same selection so the operator can paste without confusion."""
+    canon = ",".join(str(i) for i in sorted(image_ids))
+    digest = hashlib.sha256(canon.encode("utf-8")).hexdigest()
+    return digest[:8]
+
+
+@admin_bp.route("/artists/<slug>/cascade-delete", methods=["POST"])
+async def artist_cascade_delete(slug: str):
+    body = await request.get_json(silent=True) or {}
+    dry_run = bool(body.get("dry_run", False))
+    supplied_confirm = body.get("confirm", "")
+
+    async with get_session() as session:
+        artist = (await session.execute(
+            select(Artist).where(Artist.slug == slug)
+        )).scalar_one_or_none()
+        if artist is None:
+            return _bad("not_found", status=404)
+        artist_id = artist.id
+
+        projected = await session.run_sync(
+            lambda sync_sess: project_artist_cascade(sync_sess, slug=slug)
+        )
+
+    if dry_run:
+        return jsonify(projected)
+
+    expected = f"delete-artist-{artist_id}"
+    if supplied_confirm != expected:
+        return _bad(
+            "confirm_mismatch",
+            detail=f"confirm must equal {expected!r}",
+            expected=expected,
+        )
+
+    from ..tasks.admin import delete_artist_cascade_task
+    async_result = delete_artist_cascade_task.delay(artist_id=artist_id)
+    return jsonify({"task_id": async_result.id}), 202
+
+
+@admin_bp.route("/images/bulk-delete", methods=["POST"])
+async def images_bulk_delete():
+    body = await request.get_json(silent=True) or {}
+    image_ids = body.get("image_ids")
+    if not isinstance(image_ids, list) or not image_ids:
+        return _bad("invalid_image_ids", detail="image_ids must be non-empty list of int")
+    try:
+        image_ids = [int(i) for i in image_ids]
+    except (TypeError, ValueError):
+        return _bad("invalid_image_ids", detail="image_ids must contain only ints")
+
+    dry_run = bool(body.get("dry_run", False))
+    supplied_confirm = body.get("confirm", "")
+
+    async with get_session() as session:
+        projected = await session.run_sync(
+            lambda sync_sess: project_bulk_image_delete(
+                sync_sess, image_ids=image_ids,
+            )
+        )
+
+    if dry_run:
+        return jsonify(projected)
+
+    sha8 = _bulk_image_confirm_token(image_ids)
+    expected = f"delete-images-{sha8}"
+    if supplied_confirm != expected:
+        return _bad(
+            "confirm_mismatch",
+            detail=f"confirm must equal {expected!r}",
+            expected=expected,
+        )
+
+    from ..tasks.admin import bulk_delete_images_task
+    async_result = bulk_delete_images_task.delay(image_ids=image_ids)
+    return jsonify({"task_id": async_result.id}), 202
+
+
+@admin_bp.route("/tags/<int:tag_id>", methods=["DELETE"])
+async def tag_delete(tag_id: int):
+    """Tier-B sync delete. UI yes/no modal is the only confirmation."""
+    from ..services.cleanup_service import delete_tag
+
+    async with get_session() as session:
+        try:
+            result = await session.run_sync(
+                lambda sync_sess: delete_tag(sync_sess, tag_id=tag_id)
+            )
+        except LookupError:
+            return _bad("not_found", status=404)
+    return jsonify(result)
+
+
+@admin_bp.route("/tags/<int:dest_id>/merge", methods=["POST"])
+async def tag_merge(dest_id: int):
+    """Wraps TagService.merge. Source repoints to dest, dest survives,
+    source row deleted, protective alias auto-created if source was
+    ML-applied or allowlisted."""
+    from ..services.tag_service import TagMergeConflict, TagService, TagValidationError
+
+    body = await request.get_json(silent=True) or {}
+    source_id = body.get("source_id")
+    if not isinstance(source_id, int) or source_id == dest_id:
+        return _bad("invalid_source_id", detail="source_id must be int and differ from dest")
+
+    async with get_session() as session:
+        try:
+            result = await TagService(session).merge(
+                source_id=source_id, target_id=dest_id,
+            )
+        except TagMergeConflict as exc:
+            return _bad("merge_conflict", status=409, detail=str(exc))
+        except TagValidationError as exc:
+            return _bad("tag_kind_mismatch", detail=str(exc))
+        except LookupError:
+            return _bad("not_found", status=404)
+
+    # MergeResult is a frozen dataclass — flatten to dict.
+    return jsonify({
+        "result": {
+            "target_id": result.target_id,
+            "target_name": result.target_name,
+            "target_kind": result.target_kind,
+            "merged_count": result.merged_count,
+            "alias_created": result.alias_created,
+            "source_deleted": result.source_deleted,
+        },
+    })
+
+
+@admin_bp.route("/tags/<int:tag_id>/usage-count", methods=["GET"])
+async def tag_usage_count(tag_id: int):
+    """Helper for the Tier-B yes/no prompt; surfaces "N associations"
+    in the dialog so the operator knows what they're nuking."""
+    from ..services.cleanup_service import count_tag_associations
+
+    async with get_session() as session:
+        count = await session.run_sync(
+            lambda sync_sess: count_tag_associations(
+                sync_sess, tag_id=tag_id,
+            )
+        )
+    return jsonify({"count": count})
+
+
+@admin_bp.route("/tags/prune-unused", methods=["POST"])
+async def tags_prune_unused():
+    """Tier-A: dry-run preview list IS the prompt. UI calls with
+    dry_run=true first, shows the list, operator clicks button to
+    re-call with dry_run=false."""
+    from ..services.cleanup_service import prune_unused_tags
+
+    body = await request.get_json(silent=True) or {}
+    dry_run = bool(body.get("dry_run", False))
+
+    async with get_session() as session:
+        result = await session.run_sync(
+            lambda sync_sess: prune_unused_tags(
+                sync_sess, dry_run=dry_run,
+            )
+        )
+    return jsonify(result)

backend/app/api/migrate.py

@@ -1,14 +1,11 @@
 """FC-5: /api/migrate — trigger and poll migration runs.
 
 Ingest kinds (gs_ingest, ir_ingest) accept multipart/form-data with an
-`export_file` field. All other kinds accept JSON. Apply-without-backup
+`export_file` field. All other kinds accept JSON. Backup + rollback
-guard rejects non-dry-run ingests unless a pre_migration-tagged backup
+were retired in FC-3h (2026-05-24); use /api/system/backup/* instead.
-exists in the last 24h (override with body.force=true).
 """
 
 import json
-from datetime import UTC, datetime, timedelta
-from pathlib import Path
 
 from quart import Blueprint, jsonify, request
 from sqlalchemy import select
@@ -19,12 +16,12 @@ from ..tasks.migration import run_migration
 
 migrate_bp = Blueprint("migrate", __name__, url_prefix="/api/migrate")
 
+# 'backup' + 'rollback' retired 2026-05-24 (FC-3h); see /api/system/backup/*.
 _VALID_KINDS = frozenset({
-    "backup", "gs_ingest", "ir_ingest", "tag_apply",
+    "gs_ingest", "ir_ingest", "tag_apply",
-    "ml_queue", "verify", "rollback", "cleanup",
+    "ml_queue", "verify", "cleanup",
 })
 _INGEST_KINDS = frozenset({"gs_ingest", "ir_ingest"})
-_APPLY_KINDS = frozenset({"gs_ingest", "ir_ingest", "tag_apply", "rollback", "cleanup"})
 
 
 def _bad(error: str, *, status: int = 400, **extra):
@@ -33,19 +30,6 @@ def _bad(error: str, *, status: int = 400, **extra):
     return jsonify(body), status
 
 
-def _has_recent_pre_migration_backup() -> bool:
-    from ..services.migrators import backup as backup_mod
-    images_root = Path("/images")
-    manifest = backup_mod.find_latest_backup(images_root, tag="pre_migration")
-    if manifest is None:
-        return False
-    created_at_str = manifest.get("created_at")
-    if not created_at_str:
-        return False
-    created_at = datetime.fromisoformat(created_at_str)
-    return (datetime.now(UTC) - created_at) < timedelta(hours=24)
-
-
 def _run_to_dict(run: MigrationRun) -> dict:
     return {
         "id": run.id,
@@ -78,7 +62,6 @@ async def create_run(kind: str):
         except (UnicodeDecodeError, json.JSONDecodeError) as exc:
             return _bad("invalid_export_file", detail=str(exc))
         dry_run = str(form.get("dry_run", "false")).lower() in ("true", "1", "yes")
-        force = str(form.get("force", "false")).lower() in ("true", "1", "yes")
         params: dict = {"data": data, "dry_run": dry_run}
     else:
         body = await request.get_json()
@@ -87,20 +70,8 @@ async def create_run(kind: str):
         if not isinstance(body, dict):
             return _bad("invalid_body")
         dry_run = bool(body.get("dry_run", False))
-        force = bool(body.get("force", False))
         params = dict(body)
 
-    is_apply = (kind in _APPLY_KINDS) and not dry_run
-    if is_apply and not force and not _has_recent_pre_migration_backup():
-        return _bad(
-            "no_backup",
-            detail="apply action requires a pre_migration-tagged backup "
-                   "in the last 24h (or force=true).",
-        )
-
-    if kind == "backup":
-        params.setdefault("tag", "pre_migration")
-
     async with get_session() as session:
         run = MigrationRun(kind=kind, status="pending", dry_run=dry_run)
         session.add(run)

backend/app/api/system_activity.py

@@ -0,0 +1,209 @@
+"""FC-3i: system activity dashboard endpoints.
+
+Read-only. Combines Redis-broker queue depths (LLEN per queue),
+Celery worker introspection (celery inspect), and the task_run DB
+history into the surfaces the SystemActivityTab UI consumes.
+
+All filesystem/sync-client work goes through asyncio.to_thread per
+ASYNC230/240 (mirrors backend.app.api.extension's pattern).
+"""
+
+from __future__ import annotations
+
+import asyncio
+import time
+from datetime import UTC, datetime, timedelta
+
+from quart import Blueprint, jsonify, request
+from sqlalchemy import desc, func, select
+
+from ..config import get_config
+from ..extensions import get_session
+from ..models import TaskRun
+
+system_activity_bp = Blueprint(
+    "system_activity", __name__, url_prefix="/api/system/activity",
+)
+
+# Canonical queue order — must match celery_app.task_routes. UI renders
+# in this order; queues with no LLEN response show as null rather than
+# absent.
+_QUEUE_NAMES = (
+    "default", "import", "thumbnail", "ml",
+    "download", "scan", "maintenance",
+)
+
+# Cache module-level so all requests share the cache between polls.
+# Tests can reset via direct dict mutation if needed.
+_QUEUE_CACHE: dict = {"ts": 0.0, "data": None}
+_WORKER_CACHE: dict = {"ts": 0.0, "data": None}
+_QUEUE_CACHE_TTL = 2.0
+_WORKER_CACHE_TTL = 5.0
+
+
+def _read_queues_sync() -> dict:
+    """Reads each queue's LLEN from the broker. Sync — caller wraps in
+    asyncio.to_thread. Per-queue try/except returns None on failure so
+    one bad queue doesn't break the whole response."""
+    import redis  # local import; only this endpoint needs it
+
+    cfg = get_config()
+    client = redis.Redis.from_url(cfg.celery_broker_url)
+    out: dict = {}
+    for name in _QUEUE_NAMES:
+        try:
+            out[name] = int(client.llen(name))
+        except Exception:  # noqa: BLE001 — broker hiccup shouldn't break UI
+            out[name] = None
+    return {
+        "queues": out,
+        "fetched_at": datetime.now(UTC).isoformat(),
+    }
+
+
+def _read_workers_sync() -> dict:
+    """celery inspect active_queues + active. Returns per-worker info."""
+    from ..celery_app import celery as celery_app
+
+    insp = celery_app.control.inspect(timeout=2.0)
+    active_queues = insp.active_queues() or {}
+    active_tasks = insp.active() or {}
+
+    workers: dict = {}
+    for hostname, queues in active_queues.items():
+        workers[hostname] = {
+            "queues": sorted({q["name"] for q in queues}),
+            "active_count": len(active_tasks.get(hostname, [])),
+        }
+    return {
+        "workers": workers,
+        "fetched_at": datetime.now(UTC).isoformat(),
+    }
+
+
+@system_activity_bp.route("/queues", methods=["GET"])
+async def get_queues():
+    """Per-queue Redis LLEN. Cached 2s.
+
+    Response: {queues: {name: depth_or_null}, fetched_at: iso8601}
+    """
+    now = time.time()
+    if _QUEUE_CACHE["data"] is None or (now - _QUEUE_CACHE["ts"]) > _QUEUE_CACHE_TTL:
+        _QUEUE_CACHE["data"] = await asyncio.to_thread(_read_queues_sync)
+        _QUEUE_CACHE["ts"] = now
+    return jsonify(_QUEUE_CACHE["data"])
+
+
+@system_activity_bp.route("/workers", methods=["GET"])
+async def get_workers():
+    """Live celery inspect. Cached 5s.
+
+    Response: {workers: {hostname: {queues, active_count}}, fetched_at}
+    """
+    now = time.time()
+    if _WORKER_CACHE["data"] is None or (now - _WORKER_CACHE["ts"]) > _WORKER_CACHE_TTL:
+        _WORKER_CACHE["data"] = await asyncio.to_thread(_read_workers_sync)
+        _WORKER_CACHE["ts"] = now
+    return jsonify(_WORKER_CACHE["data"])
+
+
+@system_activity_bp.route("/runs", methods=["GET"])
+async def list_runs():
+    """Paginated task_run history. Query params:
+      queue=<name>     filter to one queue
+      status=<status>  filter to one status (running/ok/error/timeout/retry)
+      limit=<int>      default 50, max 200
+      before_id=<int>  cursor for keyset pagination
+
+    Response: {runs: [...], next_cursor: id|null}
+    """
+    try:
+        limit = min(int(request.args.get("limit", "50")), 200)
+    except ValueError:
+        return jsonify({"error": "invalid_limit"}), 400
+    if limit < 1:
+        return jsonify({"error": "invalid_limit"}), 400
+
+    queue = request.args.get("queue")
+    status = request.args.get("status")
+    before_id_raw = request.args.get("before_id")
+    before_id = int(before_id_raw) if before_id_raw else None
+
+    async with get_session() as session:
+        stmt = select(TaskRun).order_by(desc(TaskRun.id))
+        if queue:
+            stmt = stmt.where(TaskRun.queue == queue)
+        if status:
+            stmt = stmt.where(TaskRun.status == status)
+        if before_id is not None:
+            stmt = stmt.where(TaskRun.id < before_id)
+        stmt = stmt.limit(limit + 1)
+        rows = (await session.execute(stmt)).scalars().all()
+
+    has_more = len(rows) > limit
+    rows = rows[:limit]
+    return jsonify({
+        "runs": [_row_to_dict(r) for r in rows],
+        "next_cursor": rows[-1].id if has_more and rows else None,
+    })
+
+
+@system_activity_bp.route("/failures", methods=["GET"])
+async def list_failures():
+    """Recent failures across all lanes (24h window).
+
+    Response: {recent: [...], count_by_type: {ErrorClass: n}, since}
+    """
+    try:
+        limit = min(int(request.args.get("limit", "50")), 200)
+    except ValueError:
+        return jsonify({"error": "invalid_limit"}), 400
+
+    since = datetime.now(UTC) - timedelta(hours=24)
+
+    async with get_session() as session:
+        recent_stmt = (
+            select(TaskRun)
+            .where(TaskRun.status.in_(["error", "timeout"]))
+            .where(TaskRun.finished_at >= since)
+            .order_by(desc(TaskRun.finished_at))
+            .limit(limit)
+        )
+        recent = (await session.execute(recent_stmt)).scalars().all()
+
+        count_stmt = (
+            select(TaskRun.error_type, func.count(TaskRun.id))
+            .where(TaskRun.status.in_(["error", "timeout"]))
+            .where(TaskRun.finished_at >= since)
+            .group_by(TaskRun.error_type)
+            .order_by(desc(func.count(TaskRun.id)))
+        )
+        counts = (await session.execute(count_stmt)).all()
+
+    return jsonify({
+        "recent": [_row_to_dict(r) for r in recent],
+        "count_by_type": {
+            (row[0] or "Unknown"): row[1]
+            for row in counts
+        },
+        "since": since.isoformat(),
+    })
+
+
+def _row_to_dict(r: TaskRun) -> dict:
+    return {
+        "id": r.id,
+        "queue": r.queue,
+        "task_name": r.task_name,
+        "target_id": r.target_id,
+        "celery_task_id": r.celery_task_id,
+        "started_at": r.started_at.isoformat() if r.started_at else None,
+        "finished_at": r.finished_at.isoformat() if r.finished_at else None,
+        "duration_ms": r.duration_ms,
+        "status": r.status,
+        "error_type": r.error_type,
+        "error_message": r.error_message,
+        "retry_count": r.retry_count,
+        "worker_hostname": r.worker_hostname,
+        "args_summary": r.args_summary,
+    }

backend/app/api/system_backup.py

@@ -0,0 +1,264 @@
+"""FC-3h: /api/system/backup — create/list/restore/delete/tag for
+DB + image backups.
+
+Read endpoints are public on FC (operator-facing internal API; same
+posture as /api/system/activity). Write endpoints take a typed
+`confirm` body field that must match a server-generated token for
+that backup row, to prevent click-to-destroy by stale browser tabs
+or accidental cURL.
+"""
+from __future__ import annotations
+
+from quart import Blueprint, jsonify, request
+from sqlalchemy import desc, select
+
+from ..extensions import get_session
+from ..models import BackupRun, ImportSettings
+
+system_backup_bp = Blueprint(
+    "system_backup", __name__, url_prefix="/api/system/backup",
+)
+
+_KINDS = frozenset({"db", "images"})
+_TAG_MAX_LEN = 64
+_BACKUP_SETTINGS_FIELDS = (
+    "backup_db_nightly_enabled",
+    "backup_db_nightly_hour_utc",
+    "backup_db_keep_last_n",
+    "backup_images_keep_last_n",
+)
+
+
+def _bad(error: str, *, status: int = 400, **extra):
+    body = {"error": error}
+    body.update(extra)
+    return jsonify(body), status
+
+
+def _row_to_dict(r: BackupRun) -> dict:
+    return {
+        "id": r.id,
+        "kind": r.kind,
+        "status": r.status,
+        "tag": r.tag,
+        "triggered_by": r.triggered_by,
+        "started_at": r.started_at.isoformat() if r.started_at else None,
+        "finished_at": r.finished_at.isoformat() if r.finished_at else None,
+        "duration_seconds": (
+            int((r.finished_at - r.started_at).total_seconds())
+            if r.finished_at and r.started_at else None
+        ),
+        "sql_path": r.sql_path,
+        "tar_path": r.tar_path,
+        "size_bytes": r.size_bytes,
+        "error": r.error,
+        "restored_from_id": r.restored_from_id,
+        "manifest": r.manifest or {},
+    }
+
+
+def _validate_tag(tag):
+    if tag is None:
+        return None
+    if not isinstance(tag, str):
+        return _bad("invalid_tag", detail="tag must be string or null")
+    tag = tag.strip()
+    if not tag:
+        return None
+    if len(tag) > _TAG_MAX_LEN:
+        return _bad("invalid_tag", detail=f"tag too long (max {_TAG_MAX_LEN})")
+    return tag
+
+
+def _validate_backup_settings_patch(body: dict):
+    if "backup_db_nightly_enabled" in body and not isinstance(
+        body["backup_db_nightly_enabled"], bool,
+    ):
+        return _bad("invalid_value", detail="backup_db_nightly_enabled must be bool")
+    if "backup_db_nightly_hour_utc" in body:
+        v = body["backup_db_nightly_hour_utc"]
+        if not isinstance(v, int) or isinstance(v, bool) or not (0 <= v <= 23):
+            return _bad("invalid_value", detail="backup_db_nightly_hour_utc must be 0..23")
+    if "backup_db_keep_last_n" in body:
+        v = body["backup_db_keep_last_n"]
+        if not isinstance(v, int) or isinstance(v, bool) or not (1 <= v <= 365):
+            return _bad("invalid_value", detail="backup_db_keep_last_n must be 1..365")
+    if "backup_images_keep_last_n" in body:
+        v = body["backup_images_keep_last_n"]
+        if not isinstance(v, int) or isinstance(v, bool) or not (1 <= v <= 100):
+            return _bad("invalid_value", detail="backup_images_keep_last_n must be 1..100")
+    return None
+
+
+@system_backup_bp.route("/db", methods=["POST"])
+async def trigger_db_backup():
+    body = await request.get_json(silent=True) or {}
+    tag = _validate_tag(body.get("tag"))
+    if isinstance(tag, tuple):
+        return tag
+    from ..tasks.backup import backup_db_task
+    backup_db_task.delay(tag=tag, triggered_by="manual")
+    return jsonify({"status": "dispatched"}), 202
+
+
+@system_backup_bp.route("/images", methods=["POST"])
+async def trigger_images_backup():
+    body = await request.get_json(silent=True) or {}
+    tag = _validate_tag(body.get("tag"))
+    if isinstance(tag, tuple):
+        return tag
+    from ..tasks.backup import backup_images_task
+    backup_images_task.delay(tag=tag, triggered_by="manual")
+    return jsonify({"status": "dispatched"}), 202
+
+
+@system_backup_bp.route("/runs", methods=["GET"])
+async def list_runs():
+    try:
+        limit = min(int(request.args.get("limit", "50")), 200)
+    except ValueError:
+        return _bad("invalid_limit")
+    if limit < 1:
+        return _bad("invalid_limit")
+    kind = request.args.get("kind")
+    if kind is not None and kind not in _KINDS:
+        return _bad("invalid_kind", detail=f"kind must be one of {sorted(_KINDS)}")
+    before_id_raw = request.args.get("before_id")
+    before_id = int(before_id_raw) if before_id_raw else None
+
+    async with get_session() as session:
+        stmt = select(BackupRun).order_by(desc(BackupRun.id))
+        if kind:
+            stmt = stmt.where(BackupRun.kind == kind)
+        if before_id is not None:
+            stmt = stmt.where(BackupRun.id < before_id)
+        stmt = stmt.limit(limit + 1)
+        rows = (await session.execute(stmt)).scalars().all()
+
+    has_more = len(rows) > limit
+    rows = rows[:limit]
+    return jsonify({
+        "runs": [_row_to_dict(r) for r in rows],
+        "next_cursor": rows[-1].id if has_more and rows else None,
+    })
+
+
+@system_backup_bp.route("/runs/<int:run_id>", methods=["GET"])
+async def get_run(run_id: int):
+    async with get_session() as session:
+        row = await session.get(BackupRun, run_id)
+    if row is None:
+        return _bad("not_found", status=404)
+    return jsonify(_row_to_dict(row))
+
+
+@system_backup_bp.route("/runs/<int:run_id>", methods=["PATCH"])
+async def patch_run(run_id: int):
+    body = await request.get_json(silent=True) or {}
+    if "tag" not in body:
+        return _bad("invalid_body", detail="tag required")
+    tag = _validate_tag(body["tag"])
+    if isinstance(tag, tuple):
+        return tag
+
+    async with get_session() as session:
+        row = await session.get(BackupRun, run_id)
+        if row is None:
+            return _bad("not_found", status=404)
+        row.tag = tag
+        await session.commit()
+        await session.refresh(row)
+    return jsonify(_row_to_dict(row))
+
+
+@system_backup_bp.route("/runs/<int:run_id>/restore", methods=["POST"])
+async def trigger_restore(run_id: int):
+    body = await request.get_json(silent=True) or {}
+    supplied = body.get("confirm", "")
+
+    async with get_session() as session:
+        row = await session.get(BackupRun, run_id)
+        if row is None:
+            return _bad("not_found", status=404)
+        if row.status != "ok":
+            return _bad(
+                "not_restorable",
+                detail=f"source backup status={row.status!r}; only 'ok' rows are restorable",
+            )
+        expected = f"restore-{row.kind}-{row.id}"
+        if supplied != expected:
+            return _bad(
+                "confirm_mismatch",
+                detail=f"confirm must equal {expected!r}",
+                expected=expected,
+            )
+        kind = row.kind
+
+    if kind == "db":
+        from ..tasks.backup import restore_db_task
+        restore_db_task.delay(source_backup_run_id=run_id)
+    else:  # 'images' (the only other value _KINDS allows via the trigger path)
+        from ..tasks.backup import restore_images_task
+        restore_images_task.delay(source_backup_run_id=run_id)
+    return jsonify({"status": "dispatched", "kind": kind}), 202
+
+
+@system_backup_bp.route("/runs/<int:run_id>", methods=["DELETE"])
+async def delete_run(run_id: int):
+    body = await request.get_json(silent=True) or {}
+    supplied = body.get("confirm", "")
+
+    async with get_session() as session:
+        row = await session.get(BackupRun, run_id)
+        if row is None:
+            return _bad("not_found", status=404)
+        expected = f"delete-{row.kind}-{row.id}"
+        if supplied != expected:
+            return _bad(
+                "confirm_mismatch",
+                detail=f"confirm must equal {expected!r}",
+                expected=expected,
+            )
+        from ..services import backup_service
+        backup_service.unlink_artifact_files(
+            sql_path=row.sql_path, tar_path=row.tar_path,
+            manifest_path=(row.manifest or {}).get("manifest_path"),
+        )
+        await session.delete(row)
+        await session.commit()
+    return "", 204
+
+
+@system_backup_bp.route("/settings", methods=["GET"])
+async def get_settings():
+    async with get_session() as session:
+        row = (await session.execute(
+            select(ImportSettings).where(ImportSettings.id == 1)
+        )).scalar_one()
+    return jsonify({
+        "backup_db_nightly_enabled": row.backup_db_nightly_enabled,
+        "backup_db_nightly_hour_utc": row.backup_db_nightly_hour_utc,
+        "backup_db_keep_last_n": row.backup_db_keep_last_n,
+        "backup_images_keep_last_n": row.backup_images_keep_last_n,
+    })
+
+
+@system_backup_bp.route("/settings", methods=["PATCH"])
+async def patch_settings():
+    body = await request.get_json(silent=True)
+    if not isinstance(body, dict):
+        return _bad("invalid_body", detail="body must be a JSON object")
+
+    err = _validate_backup_settings_patch(body)
+    if err is not None:
+        return err
+
+    async with get_session() as session:
+        row = (await session.execute(
+            select(ImportSettings).where(ImportSettings.id == 1)
+        )).scalar_one()
+        for field in _BACKUP_SETTINGS_FIELDS:
+            if field in body:
+                setattr(row, field, body[field])
+        await session.commit()
+    return await get_settings()

backend/app/celery_app.py

@@ -31,6 +31,8 @@ def make_celery() -> Celery:
             "backend.app.tasks.migration",
             "backend.app.tasks.ml",
             "backend.app.tasks.download",
+            "backend.app.tasks.backup",
+            "backend.app.tasks.admin",
         ],
     )
     app.conf.update(
@@ -43,6 +45,8 @@ def make_celery() -> Celery:
             "backend.app.tasks.scan.*": {"queue": "scan"},
             "backend.app.tasks.maintenance.*": {"queue": "maintenance"},
             "backend.app.tasks.migration.*": {"queue": "maintenance"},
+            "backend.app.tasks.backup.*": {"queue": "maintenance"},
+            "backend.app.tasks.admin.*": {"queue": "maintenance"},
         },
         # Heavy ML tasks need fair dispatch — see ImageRepo's precedent.
         task_acks_late=True,
@@ -81,9 +85,28 @@ def make_celery() -> Celery:
                 "task": "backend.app.tasks.maintenance.cleanup_old_download_events",
                 "schedule": 86400.0,  # daily
             },
+            "recover-stalled-task-runs": {
+                "task": "backend.app.tasks.maintenance.recover_stalled_task_runs",
+                "schedule": 300.0,  # every 5 min, matches recover-interrupted-tasks
+            },
+            "prune-task-runs": {
+                "task": "backend.app.tasks.maintenance.prune_task_runs",
+                "schedule": 86400.0,  # daily
+            },
+            "fc3h-backup-db-nightly": {
+                "task": "backend.app.tasks.backup.backup_db_nightly",
+                "schedule": 3600.0,  # hourly tick; task self-gates on configured UTC hour
+            },
+            "fc3h-prune-backups": {
+                "task": "backend.app.tasks.backup.prune_backups",
+                "schedule": 86400.0,  # daily
+            },
         },
         timezone="UTC",
     )
+    # FC-3i: register task_run signal handlers (side-effect import).
+    from . import celery_signals  # noqa: F401
+
     return app
 
 

backend/app/celery_signals.py

@@ -0,0 +1,210 @@
+"""FC-3i: task_run lifecycle via Celery signals.
+
+Subscribes to task_prerun / task_postrun / task_failure / task_retry
+and persists one task_run row per task attempt. Drop-in for every
+existing and future Celery task — no per-task instrumentation.
+
+Signal handlers run inside the worker process (sync context); DB
+writes go through the existing shared sync engine
+(backend.app.tasks._sync_engine.sync_session_factory) — one engine
+per worker process, not per-task, so we don't blow Postgres
+max_connections under load (the reason FC-3g shared-engine fix
+existed).
+
+Failure-mode discipline (operator-pressed point): every handler is
+wrapped in try/except that swallows + logs. If the DB is down or the
+handler has a bug, the real task still runs — the dashboard goes
+dark for that interval. Monitoring NEVER breaks the thing it's
+monitoring.
+"""
+
+import logging
+from datetime import UTC, datetime
+
+from celery.exceptions import SoftTimeLimitExceeded
+from celery.signals import task_failure, task_postrun, task_prerun, task_retry
+
+from .models import TaskRun
+from .tasks._sync_engine import sync_session_factory
+
+log = logging.getLogger(__name__)
+
+# Celery-internal tasks that would generate dashboard noise without
+# operational value. Conservative list; extend only when a specific
+# task proves noisy.
+_UNTRACKED_TASK_NAMES = frozenset({
+    "celery.chord_unlock",
+    "celery.backend_cleanup",
+    "celery.chunks",
+})
+
+_MAX_ERROR_MESSAGE_LEN = 2000
+_MAX_ARGS_SUMMARY_LEN = 255
+_MAX_WORKER_HOSTNAME_LEN = 128
+
+# PostgreSQL Integer is signed 32-bit. Tasks called with a first-arg
+# int outside this range (e.g. an absurdly large mock value, or a
+# string-of-digits coercible to int but bigger than 2^31-1) would crash
+# the INSERT with NumericValueOutOfRange. Bound the recorded value to
+# the column's range; values outside become None (target_id is
+# nullable, so this is safe).
+_INT32_MAX = 2_147_483_647
+_INT32_MIN = -2_147_483_648
+
+
+def _queue_for(task) -> str:
+    """Reverse the task→queue routing from celery_app.task_routes.
+    Keep in sync if task_routes is reordered."""
+    name = getattr(task, "name", "") or ""
+    if name.startswith("backend.app.tasks.import_file."):
+        return "import"
+    if name.startswith("backend.app.tasks.ml."):
+        return "ml"
+    if name.startswith("backend.app.tasks.thumbnail."):
+        return "thumbnail"
+    if name.startswith("backend.app.tasks.download."):
+        return "download"
+    if name.startswith("backend.app.tasks.scan."):
+        return "scan"
+    if name.startswith((
+        "backend.app.tasks.maintenance.",
+        "backend.app.tasks.migration.",
+    )):
+        return "maintenance"
+    return "default"
+
+
+def _target_id_from_args(args) -> int | None:
+    """Best-effort: if the first positional arg parses as int AND fits
+    in the column's signed-32-bit range, record it as target_id
+    (image_id, source_id, etc.). Never raises."""
+    if not args:
+        return None
+    try:
+        value = int(args[0])
+    except (TypeError, ValueError):
+        return None
+    if value < _INT32_MIN or value > _INT32_MAX:
+        return None
+    return value
+
+
+def _truncate(s, limit: int) -> str | None:
+    if s is None:
+        return None
+    text = str(s)
+    return text if len(text) <= limit else text[:limit]
+
+
+def _is_tracked(task_name: str | None) -> bool:
+    return bool(task_name) and task_name not in _UNTRACKED_TASK_NAMES
+
+
+@task_prerun.connect
+def _on_prerun(sender=None, task_id=None, task=None, args=None,
+               kwargs=None, **_):
+    if not _is_tracked(getattr(task, "name", None)):
+        return
+    try:
+        Session = sync_session_factory()
+        with Session() as session:
+            session.add(TaskRun(
+                celery_task_id=task_id or "",
+                queue=_queue_for(task),
+                task_name=task.name,
+                target_id=_target_id_from_args(args),
+                started_at=datetime.now(UTC),
+                status="running",
+                args_summary=_truncate(repr(args), _MAX_ARGS_SUMMARY_LEN),
+                worker_hostname=_truncate(
+                    getattr(sender, "hostname", None),
+                    _MAX_WORKER_HOSTNAME_LEN,
+                ),
+            ))
+            session.commit()
+    except Exception:  # noqa: BLE001 — never break the worker
+        log.exception("task_run prerun insert failed (task=%s)",
+                      getattr(task, "name", "?"))
+
+
+def _finalize(task_id: str, *, status: str,
+              error_type: str | None = None,
+              error_message: str | None = None,
+              retry_count: int | None = None) -> None:
+    """Shared write path for postrun/failure/retry. Picks the most-
+    recent task_run row for this celery_task_id that's still 'running'
+    (retries reuse the same celery_task_id; each new attempt's prerun
+    inserts a fresh row, so finalize targets the latest running row)."""
+    try:
+        from sqlalchemy import select
+
+        Session = sync_session_factory()
+        now = datetime.now(UTC)
+        with Session() as session:
+            row = session.execute(
+                select(TaskRun)
+                .where(TaskRun.celery_task_id == task_id)
+                .where(TaskRun.status == "running")
+                .order_by(TaskRun.id.desc())
+                .limit(1)
+            ).scalar_one_or_none()
+            if row is None:
+                return  # no prerun row (untracked, insert failed, or already finalized)
+            row.finished_at = now
+            row.duration_ms = int(
+                (now - row.started_at).total_seconds() * 1000
+            )
+            row.status = status
+            if error_type is not None:
+                row.error_type = _truncate(error_type, 128)
+            if error_message is not None:
+                row.error_message = _truncate(error_message, _MAX_ERROR_MESSAGE_LEN)
+            if retry_count is not None:
+                row.retry_count = retry_count
+            session.commit()
+    except Exception:  # noqa: BLE001
+        log.exception("task_run finalize failed (task_id=%s)", task_id)
+
+
+@task_postrun.connect
+def _on_postrun(sender=None, task_id=None, task=None, args=None,
+                kwargs=None, retval=None, state=None, **_):
+    if not _is_tracked(getattr(task, "name", None)):
+        return
+    # state is one of SUCCESS/FAILURE/RETRY/etc. Only handle SUCCESS;
+    # task_failure handles FAILURE explicitly (with the exception).
+    if state != "SUCCESS":
+        return
+    _finalize(task_id, status="ok")
+
+
+@task_failure.connect
+def _on_failure(sender=None, task_id=None, exception=None,
+                args=None, kwargs=None, einfo=None, **_):
+    if not _is_tracked(getattr(sender, "name", None)):
+        return
+    status = ("timeout"
+              if isinstance(exception, SoftTimeLimitExceeded)
+              else "error")
+    _finalize(
+        task_id, status=status,
+        error_type=type(exception).__name__ if exception else "Unknown",
+        error_message=str(exception) if exception else None,
+    )
+
+
+@task_retry.connect
+def _on_retry(sender=None, request=None, reason=None, einfo=None, **_):
+    if not _is_tracked(getattr(sender, "name", None)):
+        return
+    task_id = getattr(request, "id", None)
+    if not task_id:
+        return
+    # Mark current attempt's row as 'retry' (terminal for this row).
+    # The next attempt's task_prerun inserts a fresh row.
+    _finalize(
+        task_id, status="retry",
+        error_type=type(reason).__name__ if reason else "Retry",
+        error_message=str(reason) if reason else None,
+        retry_count=getattr(request, "retries", 0),
+    )

backend/app/models/__init__.py

@@ -2,6 +2,7 @@
 
 from .app_setting import AppSetting
 from .artist import Artist
+from .backup_run import BackupRun
 from .base import Base
 from .credential import Credential
 from .download_event import DownloadEvent
@@ -21,11 +22,13 @@ from .tag_alias import TagAlias
 from .tag_allowlist import TagAllowlist
 from .tag_reference_embedding import TagReferenceEmbedding
 from .tag_suggestion_rejection import TagSuggestionRejection
+from .task_run import TaskRun
 
 __all__ = [
     "Base",
     "AppSetting",
     "Artist",
+    "BackupRun",
     "Source",
     "Credential",
     "Post",
@@ -46,4 +49,5 @@ __all__ = [
     "TagAllowlist",
     "TagReferenceEmbedding",
     "TagSuggestionRejection",
+    "TaskRun",
 ]

backend/app/models/backup_run.py

@@ -0,0 +1,55 @@
+"""FC-3h: backup_run — operator-facing artifact record for a backup run.
+
+One row per backup attempt (kind='db' or 'images'). Lifecycle
+tracking (started_at/finished_at/duration_ms/exception text) lives
+in task_run from FC-3i — this row records artifact metadata: file
+paths, sizes, tag (retention protection), and restore lineage via
+restored_from_id.
+
+Status values (String, not Postgres ENUM — per
+feedback_check_existing_enums):
+  pending   — created but task hasn't started yet (rare; usually
+              status starts as 'running' from the task body).
+  running   — backup task is in flight.
+  ok        — artifact successfully written.
+  error     — task raised; error column populated.
+  restoring — this row represents a restore attempt (kind = restored
+              kind); linked to source via restored_from_id.
+  restored  — restore completed successfully.
+"""
+
+from datetime import datetime
+
+from sqlalchemy import JSON, BigInteger, DateTime, ForeignKey, Integer, String, Text
+from sqlalchemy.orm import Mapped, mapped_column
+
+from .base import Base
+
+
+class BackupRun(Base):
+    __tablename__ = "backup_run"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True)
+    kind: Mapped[str] = mapped_column(String(16), nullable=False, index=True)
+    status: Mapped[str] = mapped_column(
+        String(16), nullable=False, default="pending", index=True,
+    )
+    tag: Mapped[str | None] = mapped_column(String(64), nullable=True, index=True)
+    triggered_by: Mapped[str] = mapped_column(String(32), nullable=False)
+    started_at: Mapped[datetime] = mapped_column(
+        DateTime(timezone=True), nullable=False, index=True,
+    )
+    finished_at: Mapped[datetime | None] = mapped_column(
+        DateTime(timezone=True), nullable=True, index=True,
+    )
+    sql_path: Mapped[str | None] = mapped_column(Text, nullable=True)
+    tar_path: Mapped[str | None] = mapped_column(Text, nullable=True)
+    size_bytes: Mapped[int | None] = mapped_column(BigInteger, nullable=True)
+    error: Mapped[str | None] = mapped_column(Text, nullable=True)
+    manifest: Mapped[dict] = mapped_column(
+        JSON, nullable=False, default=dict, server_default="{}",
+    )
+    restored_from_id: Mapped[int | None] = mapped_column(
+        ForeignKey("backup_run.id", ondelete="SET NULL"),
+        nullable=True,
+    )

backend/app/models/import_settings.py

@@ -49,3 +49,17 @@ class ImportSettings(Base):
     download_failure_warning_threshold: Mapped[int] = mapped_column(
         Integer, nullable=False, default=5
     )
+
+    # FC-3h backup knobs.
+    backup_db_nightly_enabled: Mapped[bool] = mapped_column(
+        Boolean, nullable=False, default=False,
+    )
+    backup_db_nightly_hour_utc: Mapped[int] = mapped_column(
+        Integer, nullable=False, default=3,
+    )
+    backup_db_keep_last_n: Mapped[int] = mapped_column(
+        Integer, nullable=False, default=14,
+    )
+    backup_images_keep_last_n: Mapped[int] = mapped_column(
+        Integer, nullable=False, default=3,
+    )

backend/app/models/task_run.py

@@ -0,0 +1,48 @@
+"""FC-3i: task_run — per-Celery-task lifecycle audit row.
+
+One row inserted by the task_prerun signal at task start, updated by
+task_postrun / task_failure / task_retry. The shape supports the
+SystemActivity dashboard's three panes: per-queue queue+worker summary,
+recent failures (24h, grouped by error_type), and full paginated
+activity history.
+
+Retention: ok rows pruned after 24h, error/timeout after 7d (see
+backend.app.tasks.maintenance.prune_task_runs).
+
+Recovery: rows stuck in 'running' for >5 min flipped to 'error' by
+backend.app.tasks.maintenance.recover_stalled_task_runs (Beat 5 min).
+"""
+
+from datetime import datetime
+
+from sqlalchemy import DateTime, Integer, String, Text
+from sqlalchemy.orm import Mapped, mapped_column
+
+from .base import Base
+
+
+class TaskRun(Base):
+    __tablename__ = "task_run"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True)
+    celery_task_id: Mapped[str] = mapped_column(
+        String(64), nullable=False, index=True,
+    )
+    queue: Mapped[str] = mapped_column(String(32), nullable=False, index=True)
+    task_name: Mapped[str] = mapped_column(String(128), nullable=False, index=True)
+    target_id: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    started_at: Mapped[datetime] = mapped_column(
+        DateTime(timezone=True), nullable=False, index=True,
+    )
+    finished_at: Mapped[datetime | None] = mapped_column(
+        DateTime(timezone=True), nullable=True, index=True,
+    )
+    duration_ms: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    status: Mapped[str] = mapped_column(
+        String(16), nullable=False, default="running", index=True,
+    )
+    error_type: Mapped[str | None] = mapped_column(String(128), nullable=True)
+    error_message: Mapped[str | None] = mapped_column(Text, nullable=True)
+    retry_count: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    worker_hostname: Mapped[str | None] = mapped_column(String(128), nullable=True)
+    args_summary: Mapped[str | None] = mapped_column(String(255), nullable=True)

backend/app/services/backup_service.py

@@ -0,0 +1,196 @@
+"""FC-3h: first-class backup/restore service for FC.
+
+Two independent backup kinds:
+  - 'db'     — pg_dump only; fast; nightly via Beat (settings-gated)
+  - 'images' — tar+zstd of /images; slow; manual trigger only
+
+Files live under <images_root>/_backups/. Each backup writes:
+  fc_<kind>_<ts>.{sql|tar.zst}        — the artifact
+  fc_<kind>_<ts>.json                 — manifest (kind/tag/triggered_by)
+
+Service functions are sync (subprocess-bound). Celery tasks in
+backend.app.tasks.backup wrap each one with task_run-tracked
+lifecycle + soft/hard time limits + retention bookkeeping.
+"""
+from __future__ import annotations
+
+import json
+import subprocess
+from datetime import UTC, datetime
+from pathlib import Path
+
+_BACKUPS_DIRNAME = "_backups"
+
+# Subprocess-level guardrails BEYOND the Celery soft_time_limit. The
+# Celery soft limit signals the Python process; subprocess.Popen in a
+# blocking syscall ignores that signal. These bound the worst case.
+_DB_SUBPROCESS_TIMEOUT_S = 12 * 60        # 12 min (Celery soft is 10 min)
+_IMAGES_SUBPROCESS_TIMEOUT_S = 7 * 60 * 60  # 7 hr (Celery soft is 6 hr)
+
+
+def _libpq_url(sa_url: str) -> str:
+    """Strip SQLAlchemy +psycopg/+asyncpg driver suffix for pg_dump/psql."""
+    for driver in (
+        "postgresql+psycopg",
+        "postgresql+asyncpg",
+        "postgresql+psycopg2",
+    ):
+        if sa_url.startswith(driver + "://"):
+            return "postgresql://" + sa_url[len(driver) + 3:]
+    return sa_url
+
+
+def _backups_dir(images_root: Path) -> Path:
+    p = images_root / _BACKUPS_DIRNAME
+    p.mkdir(parents=True, exist_ok=True)
+    return p
+
+
+def _now_ts() -> str:
+    return datetime.now(UTC).strftime("%Y%m%dT%H%M%SZ")
+
+
+def _file_size_or_none(path: Path) -> int | None:
+    try:
+        return path.stat().st_size
+    except OSError:
+        return None
+
+
+def _write_manifest(
+    out_dir: Path, *, kind: str, ts: str,
+    tag: str | None, triggered_by: str,
+    artifact_path: Path,
+) -> Path:
+    manifest = {
+        "kind": kind,
+        "backup_id": f"fc_{kind}_{ts}",
+        "tag": tag,
+        "triggered_by": triggered_by,
+        "created_at": datetime.now(UTC).isoformat(),
+        "artifact_path": str(artifact_path),
+    }
+    mf = out_dir / f"fc_{kind}_{ts}.json"
+    mf.write_text(json.dumps(manifest, indent=2))
+    return mf
+
+
+def backup_db(
+    *, db_url: str, images_root: Path,
+    tag: str | None = None, triggered_by: str = "manual",
+) -> dict:
+    """Run pg_dump; write .sql + manifest; return dict for the caller
+    to persist into BackupRun. Raises on subprocess failure."""
+    ts = _now_ts()
+    out_dir = _backups_dir(images_root)
+    sql_path = out_dir / f"fc_db_{ts}.sql"
+    subprocess.run(
+        [
+            "pg_dump", "--no-owner", "--no-acl",
+            "-f", str(sql_path), _libpq_url(db_url),
+        ],
+        capture_output=True, check=True,
+        timeout=_DB_SUBPROCESS_TIMEOUT_S,
+    )
+    manifest_path = _write_manifest(
+        out_dir, kind="db", ts=ts, tag=tag, triggered_by=triggered_by,
+        artifact_path=sql_path,
+    )
+    return {
+        "kind": "db",
+        "ts": ts,
+        "sql_path": str(sql_path),
+        "tar_path": None,
+        "manifest_path": str(manifest_path),
+        "size_bytes": _file_size_or_none(sql_path),
+    }
+
+
+def backup_images(
+    *, images_root: Path,
+    tag: str | None = None, triggered_by: str = "manual",
+) -> dict:
+    """Run tar --zstd over images_root; write .tar.zst + manifest."""
+    ts = _now_ts()
+    out_dir = _backups_dir(images_root)
+    tar_path = out_dir / f"fc_images_{ts}.tar.zst"
+    subprocess.run(
+        [
+            "tar", "--zstd", "-cf", str(tar_path),
+            "-C", str(images_root.parent), images_root.name,
+            f"--exclude={images_root.name}/_backups",
+            f"--exclude={images_root.name}/_quarantine",
+        ],
+        capture_output=True, check=True,
+        timeout=_IMAGES_SUBPROCESS_TIMEOUT_S,
+    )
+    manifest_path = _write_manifest(
+        out_dir, kind="images", ts=ts, tag=tag, triggered_by=triggered_by,
+        artifact_path=tar_path,
+    )
+    return {
+        "kind": "images",
+        "ts": ts,
+        "sql_path": None,
+        "tar_path": str(tar_path),
+        "manifest_path": str(manifest_path),
+        "size_bytes": _file_size_or_none(tar_path),
+    }
+
+
+def restore_db(*, db_url: str, sql_path: Path) -> None:
+    """Wipe public schema, then load from .sql. Raises on subprocess
+    failure; partial-restore state is the caller's concern."""
+    libpq = _libpq_url(db_url)
+    subprocess.run(
+        [
+            "psql", libpq, "-c",
+            "DROP SCHEMA IF EXISTS public CASCADE; CREATE SCHEMA public;",
+        ],
+        capture_output=True, check=True, timeout=120,
+    )
+    subprocess.run(
+        ["psql", libpq, "-f", str(sql_path)],
+        capture_output=True, check=True,
+        timeout=_DB_SUBPROCESS_TIMEOUT_S,
+    )
+
+
+def restore_images(*, images_root: Path, tar_path: Path) -> None:
+    """Untar over images_root.parent. Additive — files NOT in the
+    tarball are NOT removed. Caller wipes first if a clean restore
+    is needed."""
+    subprocess.run(
+        [
+            "tar", "--zstd", "-xf", str(tar_path),
+            "-C", str(images_root.parent),
+        ],
+        capture_output=True, check=True,
+        timeout=_IMAGES_SUBPROCESS_TIMEOUT_S,
+    )
+
+
+def unlink_artifact_files(
+    *,
+    sql_path: str | None,
+    tar_path: str | None,
+    manifest_path: str | None,
+) -> dict:
+    """Best-effort unlink of all on-disk files for a BackupRun row.
+    Returns dict keyed by label with True/False per file. Missing
+    files count as success (missing_ok semantics)."""
+    deleted: dict = {}
+    for label, p in (
+        ("sql", sql_path),
+        ("tar", tar_path),
+        ("manifest", manifest_path),
+    ):
+        if not p:
+            continue
+        path = Path(p)
+        try:
+            path.unlink(missing_ok=True)
+            deleted[label] = True
+        except OSError:
+            deleted[label] = False
+    return deleted

backend/app/services/cleanup_service.py

@@ -0,0 +1,367 @@
+"""FC-3k: first-class admin destructive operations.
+
+Projections are pure SELECTs used by both dry-run preview endpoints
+and Tier-B count prompts. Mutations (Task 2) are called from sync
+HTTP handlers (small ops) and from Celery tasks in
+backend.app.tasks.admin (long ops).
+
+This module is the PERMANENT home of artist-cascade + image-unlink
+logic. The legacy copy at backend/app/services/migrators/cleanup.py
+stays in place until FC-3j; FC-3j will replace its body with thin
+re-exports from this module and then delete the wrapper.
+"""
+from __future__ import annotations
+
+from pathlib import Path
+
+from sqlalchemy import func, select
+from sqlalchemy.orm import Session
+
+from ..models import Artist, ImageRecord, Tag
+from ..models.series_page import SeriesPage
+from ..models.tag import image_tag
+
+
+def project_artist_cascade(session: Session, *, slug: str) -> dict:
+    """Read-only projection of what delete_artist_cascade would touch.
+
+    Returns:
+        {
+          "artist": {"id": int, "name": str, "slug": str},
+          "projected": {
+            "images": int,
+            "sources": int,
+            "thumbs": int,        # images with a thumbnail_path set
+            "import_tasks": int,  # ImportTask rows referencing the artist's images
+            "bytes_on_disk": int, # SUM(image_record.size_bytes) — column is NOT NULL
+          },
+        }
+    Raises LookupError if slug not found. No mutations.
+    """
+    from ..models.import_task import ImportTask
+    from ..models.source import Source
+
+    artist = session.execute(
+        select(Artist).where(Artist.slug == slug)
+    ).scalar_one_or_none()
+    if artist is None:
+        raise LookupError(f"artist slug not found: {slug!r}")
+
+    images_count = session.execute(
+        select(func.count(ImageRecord.id))
+        .where(ImageRecord.artist_id == artist.id)
+    ).scalar_one()
+    sources_count = session.execute(
+        select(func.count(Source.id))
+        .where(Source.artist_id == artist.id)
+    ).scalar_one()
+    thumbs_count = session.execute(
+        select(func.count(ImageRecord.id))
+        .where(ImageRecord.artist_id == artist.id)
+        .where(ImageRecord.thumbnail_path.is_not(None))
+    ).scalar_one()
+    import_tasks_count = session.execute(
+        select(func.count(ImportTask.id))
+        .where(
+            ImportTask.result_image_id.in_(
+                select(ImageRecord.id).where(ImageRecord.artist_id == artist.id)
+            )
+        )
+    ).scalar_one()
+    bytes_on_disk = session.execute(
+        select(func.coalesce(func.sum(ImageRecord.size_bytes), 0))
+        .where(ImageRecord.artist_id == artist.id)
+    ).scalar_one()
+
+    return {
+        "artist": {"id": artist.id, "name": artist.name, "slug": artist.slug},
+        "projected": {
+            "images": images_count,
+            "sources": sources_count,
+            "thumbs": thumbs_count,
+            "import_tasks": import_tasks_count,
+            "bytes_on_disk": int(bytes_on_disk),
+        },
+    }
+
+
+def project_bulk_image_delete(
+    session: Session, *, image_ids: list[int],
+) -> dict:
+    """Read-only projection of what delete_images would touch.
+
+    Returns:
+        {
+          "images_found": int,
+          "thumbs_to_unlink": int,
+          "bytes_on_disk": int,
+          "missing_ids": list[int],   # ids passed in that don't exist
+        }
+    No mutations.
+    """
+    if not image_ids:
+        return {
+            "images_found": 0,
+            "thumbs_to_unlink": 0,
+            "bytes_on_disk": 0,
+            "missing_ids": [],
+        }
+
+    rows = session.execute(
+        select(
+            ImageRecord.id,
+            ImageRecord.thumbnail_path,
+            ImageRecord.size_bytes,
+        ).where(ImageRecord.id.in_(image_ids))
+    ).all()
+    found_ids = {r.id for r in rows}
+    missing = sorted(set(image_ids) - found_ids)
+    return {
+        "images_found": len(rows),
+        "thumbs_to_unlink": sum(1 for r in rows if r.thumbnail_path),
+        "bytes_on_disk": sum(r.size_bytes for r in rows),
+        "missing_ids": missing,
+    }
+
+
+def count_tag_associations(session: Session, *, tag_id: int) -> int:
+    """COUNT(*) FROM image_tag WHERE tag_id=?. For Tier-B prompt."""
+    return session.execute(
+        select(func.count())
+        .select_from(image_tag)
+        .where(image_tag.c.tag_id == tag_id)
+    ).scalar_one()
+
+
+def find_unused_tags(
+    session: Session, *, limit: int | None = None,
+) -> list[Tag]:
+    """Tags with no image_tag rows AND no series_page rows.
+
+    Sorted by name. Used by both dry-run preview and the live prune.
+    A tag is "unused" iff it has zero rows in image_tag AND zero rows
+    in series_page (so we don't accidentally prune a series tag that
+    happens to have no images yet).
+    """
+    used_via_image_tag = select(image_tag.c.tag_id).distinct()
+    used_via_series = select(SeriesPage.series_tag_id).where(
+        SeriesPage.series_tag_id.is_not(None)
+    ).distinct()
+    stmt = (
+        select(Tag)
+        .where(Tag.id.not_in(used_via_image_tag))
+        .where(Tag.id.not_in(used_via_series))
+        .order_by(Tag.name)
+    )
+    if limit is not None:
+        stmt = stmt.limit(limit)
+    return list(session.execute(stmt).scalars().all())
+
+
+def unlink_image_files(
+    image: ImageRecord, images_root: Path,
+) -> dict:
+    """Best-effort unlink of all on-disk files for an ImageRecord.
+
+    Targets: image.path (original), image.thumbnail_path (cached
+    thumbnail), and the computed thumbs path at
+    /images/thumbs/<sha256[:3]>/<sha256>.(jpg|png|webp) (tries all
+    three extensions; missing extension is silently OK).
+
+    Returns {"original": bool, "thumbnail": bool}. Missing files
+    count as success (missing_ok semantics). OSErrors are swallowed
+    and reported as False so the calling DB delete still proceeds.
+    """
+    out = {"original": False, "thumbnail": False}
+    if image.path:
+        try:
+            Path(image.path).unlink(missing_ok=True)
+            out["original"] = True
+        except OSError:
+            out["original"] = False
+    # Custom thumbnail_path (when set) — try it first.
+    if image.thumbnail_path:
+        try:
+            Path(image.thumbnail_path).unlink(missing_ok=True)
+            out["thumbnail"] = True
+        except OSError:
+            out["thumbnail"] = False
+    # Convention thumbs dir — try all extensions; missing OK.
+    if image.sha256:
+        bucket = image.sha256[:3]
+        for ext in ("jpg", "png", "webp"):
+            try:
+                (images_root / "thumbs" / bucket / f"{image.sha256}.{ext}").unlink(
+                    missing_ok=True,
+                )
+            except OSError:
+                pass
+    return out
+
+
+def delete_artist_cascade(
+    session: Session, *, artist_id: int, images_root: Path,
+) -> dict:
+    """Batched delete of an artist's images + the artist row.
+
+    Mirrors the cleanup_artist_async pattern: 500-row batches,
+    commit between batches so partial progress survives a worker
+    kill. Idempotent on missing artist (returns zeroed counts).
+    Postgres cascades handle image_tag / image_provenance /
+    series_page / tag_suggestion_rejection from ImageRecord delete,
+    and source / post / download_event / etc. from Artist delete
+    (via Artist.sources cascade="all, delete-orphan").
+    """
+    artist = session.get(Artist, artist_id)
+    if artist is None:
+        return {
+            "artist": None,
+            "summary": {
+                "images_deleted": 0,
+                "files_deleted": 0,
+                "thumbs_deleted": 0,
+                "import_tasks_nulled": 0,
+                "files_failed": 0,
+            },
+        }
+    artist_info = {"id": artist.id, "name": artist.name, "slug": artist.slug}
+
+    images_deleted = 0
+    files_deleted = 0
+    thumbs_deleted = 0
+    files_failed = 0
+
+    while True:
+        rows = session.execute(
+            select(ImageRecord)
+            .where(ImageRecord.artist_id == artist.id)
+            .limit(500)
+        ).scalars().all()
+        if not rows:
+            break
+        for img in rows:
+            unlinked = unlink_image_files(img, images_root)
+            if unlinked["original"]:
+                files_deleted += 1
+            else:
+                files_failed += 1
+            if unlinked["thumbnail"]:
+                thumbs_deleted += 1
+            session.delete(img)
+            images_deleted += 1
+        session.commit()
+
+    # ImportTask.result_image_id FK is SET NULL on image delete (Postgres
+    # handles this in the cascade above). We don't separately count those
+    # in FC-3k — the legacy cleanup_artist_async did it via
+    # source_path_prefix matching that's out of scope here.
+    import_tasks_nulled = 0
+
+    session.delete(artist)
+    session.commit()
+
+    return {
+        "artist": artist_info,
+        "summary": {
+            "images_deleted": images_deleted,
+            "files_deleted": files_deleted,
+            "thumbs_deleted": thumbs_deleted,
+            "import_tasks_nulled": import_tasks_nulled,
+            "files_failed": files_failed,
+        },
+    }
+
+
+def delete_images(
+    session: Session, *, image_ids: list[int], images_root: Path,
+) -> dict:
+    """Delete a list of images in 500-row batches with commit between.
+
+    Postgres CASCADE on image_tag / image_provenance / series_page /
+    tag_suggestion_rejection / post_attachment(FK SET NULL) handles
+    the DB side; this function handles file unlinks first then row
+    deletes. Idempotent on missing IDs (returned as missing_ids;
+    no error). On partial OSError, the row is still deleted and
+    files_failed is incremented.
+    """
+    if not image_ids:
+        return {
+            "images_deleted": 0,
+            "files_deleted": 0,
+            "thumbs_deleted": 0,
+            "files_failed": 0,
+            "missing_ids": [],
+        }
+
+    seen_ids: set[int] = set()
+    images_deleted = 0
+    files_deleted = 0
+    thumbs_deleted = 0
+    files_failed = 0
+
+    pending = list(image_ids)
+    while pending:
+        batch_ids = pending[:500]
+        pending = pending[500:]
+        rows = session.execute(
+            select(ImageRecord).where(ImageRecord.id.in_(batch_ids))
+        ).scalars().all()
+        for img in rows:
+            seen_ids.add(img.id)
+            unlinked = unlink_image_files(img, images_root)
+            if unlinked["original"]:
+                files_deleted += 1
+            else:
+                files_failed += 1
+            if unlinked["thumbnail"]:
+                thumbs_deleted += 1
+            session.delete(img)
+            images_deleted += 1
+        session.commit()
+
+    missing = sorted(set(image_ids) - seen_ids)
+    return {
+        "images_deleted": images_deleted,
+        "files_deleted": files_deleted,
+        "thumbs_deleted": thumbs_deleted,
+        "files_failed": files_failed,
+        "missing_ids": missing,
+    }
+
+
+def delete_tag(session: Session, *, tag_id: int) -> dict:
+    """Simple DELETE FROM tag WHERE id=?.
+
+    Postgres cascades the rest (image_tag, tag_alias, tag_allowlist,
+    tag_reference_embedding, tag_suggestion_rejection, series_page).
+    Returns counts BEFORE delete so the caller can surface them.
+    Raises LookupError if tag_id not found.
+    """
+    tag = session.get(Tag, tag_id)
+    if tag is None:
+        raise LookupError(f"tag id not found: {tag_id}")
+    associations_count = count_tag_associations(session, tag_id=tag_id)
+    info = {"id": tag.id, "name": tag.name, "kind": tag.kind.value}
+    session.delete(tag)
+    session.commit()
+    return {"deleted": info, "associations_removed": associations_count}
+
+
+def prune_unused_tags(session: Session, *, dry_run: bool = False) -> dict:
+    """Find tags with zero references and (unless dry_run) delete them.
+
+    Returns:
+      dry_run=True:  {"count": N, "sample_names": [first 50]}
+      dry_run=False: {"deleted": N, "sample_names": [first 50]}
+    """
+    unused = find_unused_tags(session)
+    sample = [t.name for t in unused[:50]]
+    if dry_run:
+        return {"count": len(unused), "sample_names": sample}
+    ids = [t.id for t in unused]
+    if ids:
+        session.execute(
+            Tag.__table__.delete().where(Tag.id.in_(ids))
+        )
+        session.commit()
+    return {"deleted": len(ids), "sample_names": sample}

backend/app/services/migrators/__init__.py

@@ -1,6 +1,10 @@
 """FC-5 migration tooling.
 
-One module per concern (backup/rollback/gs/ir/overlap/ml_queue/verify).
+One module per concern (gs/ir/overlap/ml_queue/verify/cleanup).
 Each migrator returns a counts dict; the run_migration task wires
 that dict into MigrationRun.counts so the UI polling shows progress.
+
+backup + rollback were retired in FC-3h (2026-05-24); first-class
+backup lives at backend/app/services/backup_service.py and exposes
+its own /api/system/backup/* surface.
 """

backend/app/services/migrators/backup.py

@@ -1,146 +0,0 @@
-"""pg_dump + tar.zst-based backup, restorable via pair of subprocess calls.
-
-Backups live under <images_root>/_backups/. Each backup is two files
-(SQL + tarball) plus a manifest JSON. Tagged backups (e.g. tag='pre_migration')
-are how rollback.py finds the most recent restorable snapshot.
-"""
-from __future__ import annotations
-
-import json
-import shutil
-import subprocess
-from datetime import UTC, datetime
-from pathlib import Path
-from typing import Any
-
-_BACKUPS_DIRNAME = "_backups"
-
-
-def _libpq_url(sa_url: str) -> str:
-    """Strip SQLAlchemy driver suffix so pg_dump/psql accept the URL.
-
-    SQLAlchemy uses URLs like `postgresql+psycopg://...` or
-    `postgresql+asyncpg://...`. libpq tools (pg_dump, psql) only know
-    the plain `postgresql://` scheme.
-    """
-    for driver in ("postgresql+psycopg", "postgresql+asyncpg", "postgresql+psycopg2"):
-        if sa_url.startswith(driver + "://"):
-            return "postgresql://" + sa_url[len(driver) + 3:]
-    return sa_url
-
-
-def _backups_dir(images_root: Path | None = None) -> Path:
-    # Overridable for tests via monkeypatch.
-    root = images_root if images_root is not None else Path("/images")
-    p = root / _BACKUPS_DIRNAME
-    p.mkdir(parents=True, exist_ok=True)
-    return p
-
-
-_DEFAULT_SUBPROCESS_TIMEOUT_S = 30 * 60  # 30 minutes
-
-
-def _run_subprocess(cmd: list[str], **kwargs: Any):
-    # Overridable for tests via monkeypatch. Hard wall-clock timeout
-    # guards against pg_dump / tar / zstd hangs on NFS — without it the
-    # task pretends to be 'running' forever (operator hit this 2026-05-
-    # 23 with two backups stuck in MigrationRun). On timeout
-    # subprocess.run raises TimeoutExpired which the caller surfaces as
-    # a task error.
-    return subprocess.run(
-        cmd,
-        capture_output=True,
-        check=True,
-        timeout=_DEFAULT_SUBPROCESS_TIMEOUT_S,
-        **{k: v for k, v in kwargs.items() if not k.startswith("_")},
-    )
-
-
-def create_backup(
-    *, db_url: str, images_root: Path, tag: str = "manual",
-) -> dict:
-    """Create a backup: pg_dump SQL + tar.zst of images.
-
-    Returns a manifest dict. Writes <ts>.sql, <ts>.tar.zst, <ts>.json into
-    <images_root>/_backups/.
-    """
-    ts = datetime.now(UTC).strftime("%Y%m%dT%H%M%SZ")
-    out_dir = _backups_dir(images_root)
-    sql_path = out_dir / f"fc_{ts}.sql"
-    tar_path = out_dir / f"fc_{ts}.tar.zst"
-    manifest_path = out_dir / f"fc_{ts}.json"
-
-    _run_subprocess(
-        ["pg_dump", "--no-owner", "--no-acl", "-f", str(sql_path), _libpq_url(db_url)],
-        _test_ts=ts,
-    )
-    _run_subprocess(
-        [
-            "tar", "--zstd", "-cf", str(tar_path),
-            "-C", str(images_root.parent), images_root.name,
-            f"--exclude={images_root.name}/_backups",
-            f"--exclude={images_root.name}/_quarantine",
-        ],
-        _test_ts=ts,
-    )
-
-    manifest = {
-        "backup_id": ts,
-        "tag": tag,
-        "created_at": datetime.now(UTC).isoformat(),
-        "sql_path": str(sql_path),
-        "tar_path": str(tar_path),
-    }
-    manifest_path.write_text(json.dumps(manifest, indent=2))
-    return manifest
-
-
-def list_backups(images_root: Path) -> list[dict]:
-    out_dir = _backups_dir(images_root)
-    items = []
-    for mf in sorted(out_dir.glob("fc_*.json"), reverse=True):
-        try:
-            items.append(json.loads(mf.read_text()))
-        except Exception:
-            continue
-    return items
-
-
-def find_latest_backup(images_root: Path, *, tag: str) -> dict | None:
-    for mf in list_backups(images_root):
-        if mf.get("tag") == tag:
-            return mf
-    return None
-
-
-def restore_backup(
-    *, manifest: dict, db_url: str, images_root: Path,
-) -> dict:
-    """Restore from a backup manifest.
-
-    1. Replay the .sql via psql.
-    2. Wipe /images/ contents (except _backups/, which holds the file we're using).
-    3. Untar the .tar.zst into /images/.
-    """
-    sql_path = Path(manifest["sql_path"])
-    tar_path = Path(manifest["tar_path"])
-
-    _run_subprocess(
-        ["psql", "-d", _libpq_url(db_url), "-f", str(sql_path)],
-    )
-
-    # Wipe everything in images_root EXCEPT _backups/ (we'd delete the backup
-    # we're restoring from!).
-    for entry in images_root.iterdir():
-        if entry.name == _BACKUPS_DIRNAME:
-            continue
-        if entry.is_dir():
-            shutil.rmtree(entry)
-        else:
-            entry.unlink()
-
-    _run_subprocess(
-        ["tar", "--zstd", "-xf", str(tar_path), "-C", str(images_root.parent)],
-    )
-
-    return {"restored_from": manifest["backup_id"]}

backend/app/services/migrators/rollback.py

@@ -1,21 +0,0 @@
-"""Restore from the most recent 'pre_migration'-tagged backup."""
-from __future__ import annotations
-
-from pathlib import Path
-
-from . import backup as backup_mod
-
-
-class NoBackupFoundError(Exception):
-    """Raised when rollback() is called with no pre_migration backup on disk."""
-
-
-def rollback_to_pre_migration(*, db_url: str, images_root: Path) -> dict:
-    manifest = backup_mod.find_latest_backup(images_root, tag="pre_migration")
-    if manifest is None:
-        raise NoBackupFoundError(
-            "no pre_migration-tagged backup found under <images_root>/_backups/"
-        )
-    return backup_mod.restore_backup(
-        manifest=manifest, db_url=db_url, images_root=images_root,
-    )

backend/app/tasks/admin.py

@@ -0,0 +1,57 @@
+"""FC-3k: admin destructive Celery tasks.
+
+Two long-running ops on the maintenance queue. task_run lifecycle is
+captured automatically by FC-3i signals — these tasks just return
+their summary dict so it lands in task_run.metadata (via Celery's
+result backend) for the dashboard to surface.
+
+Soft/hard time limits inherit the FC-3i recovery sweep: a runaway
+task gets killed and flipped to status='timeout' by
+recover_stalled_task_runs.
+"""
+from __future__ import annotations
+
+import logging
+from pathlib import Path
+
+from sqlalchemy.exc import DBAPIError, OperationalError
+
+from ..celery_app import celery
+from ..services import cleanup_service
+from ._sync_engine import sync_session_factory as _sync_session_factory
+
+log = logging.getLogger(__name__)
+IMAGES_ROOT = Path("/images")
+
+
+@celery.task(
+    name="backend.app.tasks.admin.delete_artist_cascade_task",
+    bind=True,
+    autoretry_for=(OperationalError, DBAPIError),
+    retry_backoff=15, retry_backoff_max=180, max_retries=1,
+    soft_time_limit=1800, time_limit=2400,  # 30 min / 40 min
+)
+def delete_artist_cascade_task(self, *, artist_id: int) -> dict:
+    """Wraps cleanup_service.delete_artist_cascade. Returns the
+    service's summary dict for FC-3i task_run.metadata capture."""
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        return cleanup_service.delete_artist_cascade(
+            session, artist_id=artist_id, images_root=IMAGES_ROOT,
+        )
+
+
+@celery.task(
+    name="backend.app.tasks.admin.bulk_delete_images_task",
+    bind=True,
+    autoretry_for=(OperationalError, DBAPIError),
+    retry_backoff=15, retry_backoff_max=180, max_retries=1,
+    soft_time_limit=900, time_limit=1200,  # 15 min / 20 min
+)
+def bulk_delete_images_task(self, *, image_ids: list[int]) -> dict:
+    """Wraps cleanup_service.delete_images."""
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        return cleanup_service.delete_images(
+            session, image_ids=image_ids, images_root=IMAGES_ROOT,
+        )

backend/app/tasks/backup.py

@@ -0,0 +1,300 @@
+"""FC-3h: backup/restore Celery tasks.
+
+All tasks live on the maintenance queue (per celery_app.task_routes).
+task_run lifecycle tracking is automatic via FC-3i signals — these
+tasks just record the operator-facing artifact metadata into
+BackupRun.
+"""
+from __future__ import annotations
+
+import logging
+from datetime import UTC, datetime
+from pathlib import Path
+
+from celery.exceptions import SoftTimeLimitExceeded
+from sqlalchemy import select
+from sqlalchemy.exc import DBAPIError, OperationalError
+
+from ..celery_app import celery
+from ..config import get_config
+from ..models import BackupRun, ImportSettings
+from ..services import backup_service
+from ._sync_engine import sync_session_factory as _sync_session_factory
+
+log = logging.getLogger(__name__)
+IMAGES_ROOT = Path("/images")
+
+
+def _mark_failed(session, row: BackupRun, exc: BaseException) -> None:
+    """Flip a BackupRun row from running/restoring to error with a
+    truncated error message and finished_at. Caller already holds the
+    session open."""
+    row.status = "error"
+    row.error = f"{type(exc).__name__}: {exc}"[:2000]
+    row.finished_at = datetime.now(UTC)
+    session.add(row)
+    session.commit()
+
+
+@celery.task(
+    name="backend.app.tasks.backup.backup_db_task",
+    bind=True,
+    autoretry_for=(OperationalError, DBAPIError),
+    retry_backoff=10, retry_backoff_max=120, max_retries=2,
+    soft_time_limit=600, time_limit=720,
+)
+def backup_db_task(self, *, tag: str | None = None,
+                   triggered_by: str = "manual") -> dict:
+    """Create one DB backup. Returns {'backup_run_id': N}."""
+    SessionLocal = _sync_session_factory()
+    cfg = get_config()
+    now = datetime.now(UTC)
+    with SessionLocal() as session:
+        row = BackupRun(
+            kind="db", status="running", tag=tag,
+            triggered_by=triggered_by, started_at=now, manifest={},
+        )
+        session.add(row)
+        session.commit()
+        session.refresh(row)
+        run_id = row.id
+
+    try:
+        result = backup_service.backup_db(
+            db_url=cfg.database_url_sync, images_root=IMAGES_ROOT,
+            tag=tag, triggered_by=triggered_by,
+        )
+    except (SoftTimeLimitExceeded, Exception) as exc:
+        with SessionLocal() as session:
+            row = session.get(BackupRun, run_id)
+            if row is not None:
+                _mark_failed(session, row, exc)
+        raise
+
+    with SessionLocal() as session:
+        row = session.get(BackupRun, run_id)
+        row.status = "ok"
+        row.finished_at = datetime.now(UTC)
+        row.sql_path = result["sql_path"]
+        row.size_bytes = result["size_bytes"]
+        row.manifest = {
+            "manifest_path": result["manifest_path"],
+            "ts": result["ts"],
+        }
+        session.commit()
+    return {"backup_run_id": run_id}
+
+
+@celery.task(
+    name="backend.app.tasks.backup.backup_images_task",
+    bind=True,
+    autoretry_for=(OperationalError, DBAPIError),
+    retry_backoff=30, retry_backoff_max=300, max_retries=1,
+    soft_time_limit=21600, time_limit=23400,
+)
+def backup_images_task(self, *, tag: str | None = None,
+                       triggered_by: str = "manual") -> dict:
+    """Create one images backup. Same shape as backup_db_task; uses
+    tar_path instead of sql_path."""
+    SessionLocal = _sync_session_factory()
+    now = datetime.now(UTC)
+    with SessionLocal() as session:
+        row = BackupRun(
+            kind="images", status="running", tag=tag,
+            triggered_by=triggered_by, started_at=now, manifest={},
+        )
+        session.add(row)
+        session.commit()
+        session.refresh(row)
+        run_id = row.id
+
+    try:
+        result = backup_service.backup_images(
+            images_root=IMAGES_ROOT,
+            tag=tag, triggered_by=triggered_by,
+        )
+    except (SoftTimeLimitExceeded, Exception) as exc:
+        with SessionLocal() as session:
+            row = session.get(BackupRun, run_id)
+            if row is not None:
+                _mark_failed(session, row, exc)
+        raise
+
+    with SessionLocal() as session:
+        row = session.get(BackupRun, run_id)
+        row.status = "ok"
+        row.finished_at = datetime.now(UTC)
+        row.tar_path = result["tar_path"]
+        row.size_bytes = result["size_bytes"]
+        row.manifest = {
+            "manifest_path": result["manifest_path"],
+            "ts": result["ts"],
+        }
+        session.commit()
+    return {"backup_run_id": run_id}
+
+
+@celery.task(
+    name="backend.app.tasks.backup.restore_db_task",
+    bind=True,
+    max_retries=0,  # NEVER auto-retry a half-applied restore.
+    soft_time_limit=1200, time_limit=1800,
+)
+def restore_db_task(self, *, source_backup_run_id: int) -> dict:
+    """Restore from a previous DB backup. Inserts a NEW BackupRun row
+    (kind='db', status='restoring') linked to the source via
+    restored_from_id; flips to 'restored' on success or 'error' on
+    failure. Operator sees the restore as a row in the dashboard."""
+    SessionLocal = _sync_session_factory()
+    cfg = get_config()
+    now = datetime.now(UTC)
+    with SessionLocal() as session:
+        src = session.get(BackupRun, source_backup_run_id)
+        if src is None or src.kind != "db" or not src.sql_path:
+            raise ValueError(
+                f"BackupRun id={source_backup_run_id} is not a valid DB backup"
+            )
+        marker = BackupRun(
+            kind="db", status="restoring",
+            triggered_by="restore", started_at=now,
+            restored_from_id=src.id,
+            manifest={"source_sql_path": src.sql_path},
+        )
+        session.add(marker)
+        session.commit()
+        session.refresh(marker)
+        marker_id = marker.id
+        sql_path = src.sql_path
+
+    try:
+        backup_service.restore_db(
+            db_url=cfg.database_url_sync, sql_path=Path(sql_path),
+        )
+    except (SoftTimeLimitExceeded, Exception) as exc:
+        with SessionLocal() as session:
+            row = session.get(BackupRun, marker_id)
+            if row is not None:
+                _mark_failed(session, row, exc)
+        raise
+
+    with SessionLocal() as session:
+        row = session.get(BackupRun, marker_id)
+        row.status = "restored"
+        row.finished_at = datetime.now(UTC)
+        session.commit()
+    return {"backup_run_id": marker_id}
+
+
+@celery.task(
+    name="backend.app.tasks.backup.restore_images_task",
+    bind=True, max_retries=0,
+    soft_time_limit=21600, time_limit=23400,
+)
+def restore_images_task(self, *, source_backup_run_id: int) -> dict:
+    """Mirrors restore_db_task; uses backup_service.restore_images."""
+    SessionLocal = _sync_session_factory()
+    now = datetime.now(UTC)
+    with SessionLocal() as session:
+        src = session.get(BackupRun, source_backup_run_id)
+        if src is None or src.kind != "images" or not src.tar_path:
+            raise ValueError(
+                f"BackupRun id={source_backup_run_id} is not a valid images backup"
+            )
+        marker = BackupRun(
+            kind="images", status="restoring",
+            triggered_by="restore", started_at=now,
+            restored_from_id=src.id,
+            manifest={"source_tar_path": src.tar_path},
+        )
+        session.add(marker)
+        session.commit()
+        session.refresh(marker)
+        marker_id = marker.id
+        tar_path = src.tar_path
+
+    try:
+        backup_service.restore_images(
+            images_root=IMAGES_ROOT, tar_path=Path(tar_path),
+        )
+    except (SoftTimeLimitExceeded, Exception) as exc:
+        with SessionLocal() as session:
+            row = session.get(BackupRun, marker_id)
+            if row is not None:
+                _mark_failed(session, row, exc)
+        raise
+
+    with SessionLocal() as session:
+        row = session.get(BackupRun, marker_id)
+        row.status = "restored"
+        row.finished_at = datetime.now(UTC)
+        session.commit()
+    return {"backup_run_id": marker_id}
+
+
+@celery.task(
+    name="backend.app.tasks.backup.prune_backups",
+    soft_time_limit=300, time_limit=600,
+)
+def prune_backups() -> dict:
+    """Daily Beat. Per-kind retention from ImportSettings.
+
+    Returns {"db_deleted": N, "images_deleted": M, "files_unlinked": K}.
+    Tagged rows (tag IS NOT NULL) are never pruned.
+    Status='running' / 'restoring' rows are never pruned (recovery
+    sweep from FC-3i handles those via task_run).
+    """
+    SessionLocal = _sync_session_factory()
+    counts = {"db_deleted": 0, "images_deleted": 0, "files_unlinked": 0}
+    with SessionLocal() as session:
+        s = session.execute(
+            select(ImportSettings).where(ImportSettings.id == 1)
+        ).scalar_one()
+        for kind, keep in (
+            ("db", s.backup_db_keep_last_n),
+            ("images", s.backup_images_keep_last_n),
+        ):
+            candidates = session.execute(
+                select(BackupRun)
+                .where(BackupRun.kind == kind)
+                .where(BackupRun.tag.is_(None))
+                .where(BackupRun.status.in_(["ok", "error"]))
+                .order_by(BackupRun.started_at.desc())
+                .offset(keep)
+            ).scalars().all()
+            for row in candidates:
+                result = backup_service.unlink_artifact_files(
+                    sql_path=row.sql_path,
+                    tar_path=row.tar_path,
+                    manifest_path=(row.manifest or {}).get("manifest_path"),
+                )
+                counts["files_unlinked"] += sum(
+                    1 for v in result.values() if v
+                )
+                session.delete(row)
+                counts[f"{kind}_deleted"] += 1
+        session.commit()
+    return counts
+
+
+@celery.task(
+    name="backend.app.tasks.backup.backup_db_nightly",
+    soft_time_limit=60, time_limit=120,
+)
+def backup_db_nightly() -> dict:
+    """Hourly tick. Dispatches a real backup ONLY if the configured
+    UTC hour matches and the nightly setting is enabled. Returns
+    either {'skipped': '<reason>'} or {'dispatched': '<task_id>'}."""
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        s = session.execute(
+            select(ImportSettings).where(ImportSettings.id == 1)
+        ).scalar_one()
+        nightly_enabled = s.backup_db_nightly_enabled
+        configured_hour = s.backup_db_nightly_hour_utc
+    if not nightly_enabled:
+        return {"skipped": "nightly disabled"}
+    now_hour = datetime.now(UTC).hour
+    if now_hour != configured_hour:
+        return {"skipped": f"hour={now_hour} != configured={configured_hour}"}
+    res = backup_db_task.delay(triggered_by="nightly")
+    return {"dispatched": res.id}

backend/app/tasks/maintenance.py

@@ -9,7 +9,7 @@ from PIL import Image
 from sqlalchemy import delete, select, update
 
 from ..celery_app import celery
-from ..models import DownloadEvent, ImageRecord, ImportSettings, ImportTask
+from ..models import DownloadEvent, ImageRecord, ImportSettings, ImportTask, TaskRun
 from ..utils.phash import compute_phash
 from ._sync_engine import sync_session_factory as _sync_session_factory
 
@@ -20,6 +20,8 @@ OLD_TASK_DAYS = 7
 PHASH_PAGE = 500
 VERIFY_PAGE = 200
 FFPROBE_TIMEOUT_SECONDS = 10
+TASK_RUN_KEEP_OK_SECONDS = 24 * 3600       # 24 h
+TASK_RUN_KEEP_FAILURE_SECONDS = 7 * 24 * 3600  # 7 days
 
 
 @celery.task(name="backend.app.tasks.maintenance.recover_interrupted_tasks")
@@ -80,6 +82,71 @@ def cleanup_old_tasks() -> int:
         return result.rowcount or 0
 
 
+@celery.task(name="backend.app.tasks.maintenance.recover_stalled_task_runs")
+def recover_stalled_task_runs() -> int:
+    """Flip task_run rows stuck in 'running' for >STUCK_THRESHOLD_MINUTES
+    to 'error'. FC-3i.
+
+    A row gets stuck when the worker dies without emitting
+    task_postrun / task_failure (e.g. OOM, container restart between
+    signals, signal handler raised+logged). Shares the 5-min threshold
+    with recover_interrupted_tasks for consistency.
+    """
+    SessionLocal = _sync_session_factory()
+    cutoff = datetime.now(UTC) - timedelta(minutes=STUCK_THRESHOLD_MINUTES)
+    with SessionLocal() as session:
+        result = session.execute(
+            update(TaskRun)
+            .where(TaskRun.status == "running")
+            .where(TaskRun.started_at < cutoff)
+            .values(
+                status="error",
+                error_type="RecoverySweep",
+                error_message=(
+                    f"no completion signal received within "
+                    f"{STUCK_THRESHOLD_MINUTES} min"
+                ),
+                finished_at=datetime.now(UTC),
+            )
+        )
+        session.commit()
+        return result.rowcount or 0
+
+
+@celery.task(name="backend.app.tasks.maintenance.prune_task_runs")
+def prune_task_runs() -> dict:
+    """Daily retention for task_run rows. FC-3i.
+
+    - 'ok' rows: deleted after TASK_RUN_KEEP_OK_SECONDS (24h default).
+      Success is high-volume, not interesting after a day.
+    - 'error' / 'timeout' rows: deleted after TASK_RUN_KEEP_FAILURE_SECONDS
+      (7 days default). Failures are operationally interesting longer.
+    - 'running' rows: NEVER deleted by this task. The recovery sweep
+      (recover_stalled_task_runs) is the mechanism that flips them to
+      terminal state; prune doesn't touch in-flight state.
+    - 'retry' rows: treated as failures (>7d).
+
+    Returns dict of how many rows were deleted in each bucket.
+    """
+    SessionLocal = _sync_session_factory()
+    now = datetime.now(UTC)
+    ok_cutoff = now - timedelta(seconds=TASK_RUN_KEEP_OK_SECONDS)
+    fail_cutoff = now - timedelta(seconds=TASK_RUN_KEEP_FAILURE_SECONDS)
+    with SessionLocal() as session:
+        ok_deleted = session.execute(
+            delete(TaskRun)
+            .where(TaskRun.status == "ok")
+            .where(TaskRun.finished_at < ok_cutoff)
+        ).rowcount or 0
+        fail_deleted = session.execute(
+            delete(TaskRun)
+            .where(TaskRun.status.in_(["error", "timeout", "retry"]))
+            .where(TaskRun.finished_at < fail_cutoff)
+        ).rowcount or 0
+        session.commit()
+        return {"ok_deleted": ok_deleted, "failures_deleted": fail_deleted}
+
+
 @celery.task(name="backend.app.tasks.maintenance.backfill_phash")
 def backfill_phash() -> int:
     """Recompute phash for stored images that have none (imported before

backend/app/tasks/migration.py

@@ -4,7 +4,8 @@ Dispatches to the right migrator based on `kind`. Updates MigrationRun
 row's status/counts/finished_at as it runs. Failures set status='error'
 with the error message preserved.
 
-kinds: backup, gs_ingest, ir_ingest, tag_apply, ml_queue, verify, rollback, cleanup
+kinds: gs_ingest, ir_ingest, tag_apply, ml_queue, verify, cleanup
+(backup + rollback retired 2026-05-24 → see /api/system/backup/*)
 """
 from __future__ import annotations
 
@@ -20,10 +21,8 @@ from ..celery_app import celery
 from ..config import get_config
 from ..models import MigrationRun
 from ..services.credential_crypto import CredentialCrypto
-from ..services.migrators import backup as backup_mod
 from ..services.migrators import cleanup as cleanup_mod
 from ..services.migrators import gs_ingest, ir_ingest, ml_queue, tag_apply, verify
-from ..services.migrators import rollback as rollback_mod
 
 log = logging.getLogger(__name__)
 
@@ -65,21 +64,11 @@ async def _run_async(run_id: int, kind: str, params: dict) -> dict:
         async with factory() as db:
             await _update_run(db, run_id, status="running")
             try:
-                if kind == "backup":
+                if kind in ("backup", "rollback"):
-                    manifest = backup_mod.create_backup(
+                    raise ValueError(
-                        db_url=get_config().database_url_sync,
+                        f"kind {kind!r} retired in FC-3h; "
-                        images_root=IMAGES_ROOT,
+                        "use /api/system/backup/* instead"
-                        tag=params.get("tag", "manual"),
                     )
-                    await _update_run(
-                        db, run_id, status="ok",
-                        counts={"rows_processed": 0, "rows_inserted": 0,
-                                "rows_skipped": 0, "files_copied": 0,
-                                "bytes_copied": 0, "conflicts": 0},
-                        finished_at=datetime.now(UTC),
-                        metadata_patch={"manifest": manifest},
-                    )
-                    return manifest
 
                 elif kind == "gs_ingest":
                     fc_crypto = CredentialCrypto(_KEY_PATH)
@@ -166,21 +155,6 @@ async def _run_async(run_id: int, kind: str, params: dict) -> dict:
                     )
                     return result
 
-                elif kind == "rollback":
-                    result = rollback_mod.rollback_to_pre_migration(
-                        db_url=get_config().database_url_sync,
-                        images_root=IMAGES_ROOT,
-                    )
-                    await _update_run(
-                        db, run_id, status="ok",
-                        counts={"rows_processed": 0, "rows_inserted": 0,
-                                "rows_skipped": 0, "files_copied": 0,
-                                "bytes_copied": 0, "conflicts": 0},
-                        finished_at=datetime.now(UTC),
-                        metadata_patch={"rollback_result": result},
-                    )
-                    return result
-
                 else:
                     raise ValueError(f"unknown kind: {kind}")
 

frontend/src/components/artist/ArtistDangerZone.vue

@@ -0,0 +1,95 @@
+<template>
+  <v-card class="fc-danger-zone mt-8" variant="outlined">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-alert-octagon" color="error" size="small" />
+      <span>Danger zone</span>
+    </v-card-title>
+    <v-card-text>
+      <p class="text-body-2 fc-muted mb-4">
+        Cascade-delete this artist and every image, source, post, and
+        attachment associated with them. This cannot be undone.
+        Recoverable only from an FC-3h backup.
+      </p>
+      <v-btn
+        color="error" variant="flat" rounded="pill"
+        prepend-icon="mdi-delete-forever"
+        :loading="loading"
+        @click="onClick"
+      >Delete artist &amp; cascade</v-btn>
+
+      <DestructiveConfirmModal
+        v-model="modalOpen"
+        action="delete"
+        kind="artist"
+        :run-id="artistId"
+        tier="C"
+        :projected-counts="projectedCounts"
+        :description="modalDescription"
+        @confirm="onConfirm"
+      />
+    </v-card-text>
+  </v-card>
+</template>
+
+<script setup>
+import { computed, ref } from 'vue'
+import { useRouter } from 'vue-router'
+
+import DestructiveConfirmModal from '../modal/DestructiveConfirmModal.vue'
+import { useAdminStore } from '../../stores/admin.js'
+
+const props = defineProps({
+  slug: { type: String, required: true },
+  artistId: { type: Number, required: true },
+  artistName: { type: String, required: true },
+})
+
+const router = useRouter()
+const store = useAdminStore()
+const loading = ref(false)
+const modalOpen = ref(false)
+const projected = ref(null)
+
+const projectedCounts = computed(() => projected.value?.projected || null)
+
+const modalDescription = computed(
+  () => projected.value
+    ? `Artist “${props.artistName}” — `
+        + `${projected.value.projected.images} images, `
+        + `${projected.value.projected.sources} sources, `
+        + `${Math.round(projected.value.projected.bytes_on_disk / 1_048_576)} MiB on disk`
+    : '',
+)
+
+async function onClick() {
+  loading.value = true
+  try {
+    projected.value = await store.projectArtistCascade(props.slug)
+    modalOpen.value = true
+  } finally {
+    loading.value = false
+  }
+}
+
+async function onConfirm(token) {
+  loading.value = true
+  try {
+    const result = await store.dispatchArtistCascade(props.slug, token)
+    const taskId = result.task_id
+    router.push('/artists')
+    if (taskId) {
+      store.pollTaskUntilDone(taskId).catch(() => {})
+    }
+  } finally {
+    loading.value = false
+  }
+}
+</script>
+
+<style scoped>
+.fc-danger-zone {
+  border-color: rgb(var(--v-theme-error));
+  border-radius: 8px;
+}
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/components/discovery/TagCard.vue

@@ -44,7 +44,31 @@
       </div>
       <div class="fc-tagcard__meta">
         <v-chip size="x-small" label>{{ card.kind }}</v-chip>
-        <span class="fc-tagcard__count">{{ card.image_count }}</span>
+        <div class="fc-tagcard__meta-right">
+          <span class="fc-tagcard__count">{{ card.image_count }}</span>
+          <v-menu>
+            <template #activator="{ props: act }">
+              <v-btn
+                class="fc-tagcard__menu"
+                icon="mdi-dots-vertical" size="x-small" variant="text"
+                v-bind="act" @click.stop
+              />
+            </template>
+            <v-list density="compact">
+              <v-list-item
+                title="Merge with…"
+                prepend-icon="mdi-call-merge"
+                @click="$emit('merge-with', card)"
+              />
+              <v-list-item
+                title="Delete tag"
+                prepend-icon="mdi-delete"
+                base-color="error"
+                @click="$emit('delete', card)"
+              />
+            </v-list>
+          </v-menu>
+        </div>
       </div>
     </v-card-text>
   </v-card>
@@ -54,7 +78,7 @@
 import { ref } from 'vue'
 
 const props = defineProps({ card: { type: Object, required: true } })
-const emit = defineEmits(['open', 'rename', 'manage', 'read'])
+const emit = defineEmits(['open', 'rename', 'manage', 'read', 'merge-with', 'delete'])
 
 const editing = ref(false)
 const draft = ref('')
@@ -106,4 +130,13 @@ function submit() {
 }
 .fc-tagcard:hover .fc-tagcard__edit { opacity: .6; }
 .fc-tagcard__edit:hover { opacity: 1; }
+.fc-tagcard__meta-right {
+  display: flex; align-items: center; gap: 4px;
+}
+.fc-tagcard__menu {
+  opacity: 0;
+  transition: opacity .15s ease;
+}
+.fc-tagcard:hover .fc-tagcard__menu { opacity: .6; }
+.fc-tagcard__menu:hover { opacity: 1; }
 </style>

frontend/src/components/gallery/BulkEditorPanel.vue

@@ -55,19 +55,44 @@
       </template>
     </div>
 
+    <div class="fc-bulk-panel__section">
+      <h4>Destructive</h4>
+      <v-btn
+        color="error" variant="flat" rounded="pill" block
+        prepend-icon="mdi-delete-forever"
+        :disabled="!sel.count"
+        :loading="deleting"
+        @click="onDeleteClick"
+      >Delete {{ sel.count }} selected</v-btn>
+    </div>
+
     <div class="fc-bulk-panel__foot">
       <v-btn variant="text" block @click="sel.clear()">Clear selection</v-btn>
     </div>
+
+    <DestructiveConfirmModal
+      v-model="deleteModalOpen"
+      action="delete"
+      kind="images-selection"
+      :run-id="bulkToken"
+      tier="C"
+      :projected-counts="bulkProjectedCounts"
+      :description="bulkDescription"
+      @confirm="onDeleteConfirm"
+    />
   </aside>
 </template>
 
 <script setup>
-import { ref, watch } from 'vue'
+import { computed, ref, watch } from 'vue'
 import { useGallerySelectionStore } from '../../stores/gallerySelection.js'
+import { useAdminStore } from '../../stores/admin.js'
 import { useApi } from '../../composables/useApi.js'
+import DestructiveConfirmModal from '../modal/DestructiveConfirmModal.vue'
 
 const sel = useGallerySelectionStore()
 const api = useApi()
+const adminStore = useAdminStore()
 
 const addModel = ref(null)
 const addHits = ref([])
@@ -99,6 +124,66 @@ async function onAddPick(id) {
 watch(() => sel.order.length, () => {
   if (sel.isSelectMode) sel.refresh()
 })
+
+// --- FC-3k bulk delete -----------------------------------------------
+
+const deleting = ref(false)
+const deleteModalOpen = ref(false)
+const bulkProjected = ref(null)
+const bulkToken = ref('')
+
+const bulkProjectedCounts = computed(() => bulkProjected.value
+  ? {
+    images: bulkProjected.value.images_found,
+    thumbnails: bulkProjected.value.thumbs_to_unlink,
+    bytes: bulkProjected.value.bytes_on_disk,
+  }
+  : null,
+)
+const bulkDescription = computed(
+  () => bulkProjected.value
+    ? `${bulkProjected.value.images_found} images, `
+        + `${Math.round(bulkProjected.value.bytes_on_disk / 1_048_576)} MiB on disk`
+    : '',
+)
+
+async function _computeSha8(ids) {
+  const canon = [...ids].sort((a, b) => a - b).join(',')
+  const buf = new TextEncoder().encode(canon)
+  const hashBuf = await crypto.subtle.digest('SHA-256', buf)
+  const bytes = new Uint8Array(hashBuf)
+  let hex = ''
+  for (let i = 0; i < 4; i++) {
+    hex += bytes[i].toString(16).padStart(2, '0')
+  }
+  return hex
+}
+
+async function onDeleteClick() {
+  if (!sel.order.length) return
+  deleting.value = true
+  try {
+    bulkProjected.value = await adminStore.projectBulkImageDelete(sel.order)
+    bulkToken.value = await _computeSha8(sel.order)
+    deleteModalOpen.value = true
+  } finally {
+    deleting.value = false
+  }
+}
+
+async function onDeleteConfirm(token) {
+  deleting.value = true
+  try {
+    const result = await adminStore.dispatchBulkImageDelete(sel.order, token)
+    const taskId = result.task_id
+    if (taskId) {
+      adminStore.pollTaskUntilDone(taskId).catch(() => {})
+    }
+    sel.clear()
+  } finally {
+    deleting.value = false
+  }
+}
 </script>
 
 <style scoped>

frontend/src/components/modal/DestructiveConfirmModal.vue

@@ -0,0 +1,128 @@
+<template>
+  <v-dialog
+    :model-value="modelValue"
+    max-width="520" persistent
+    @update:model-value="$emit('update:modelValue', $event)"
+  >
+    <v-card>
+      <v-card-title>
+        {{ titleVerb }}
+        {{ kindLabel }}<span v-if="runId"> #{{ runId }}</span>
+      </v-card-title>
+      <v-card-text>
+        <v-alert
+          :type="alertType"
+          variant="tonal" density="compact" class="mb-3"
+        >
+          <strong>{{ warningText }}</strong>
+          <div v-if="description" class="mt-1">{{ description }}</div>
+        </v-alert>
+
+        <div v-if="projectedCounts" class="fc-counts mb-3">
+          <div v-for="(v, k) in projectedCounts" :key="k">
+            <span class="fc-counts-key">{{ k }}:</span>
+            <span class="fc-counts-val">{{ v }}</span>
+          </div>
+        </div>
+
+        <template v-if="tier === 'C'">
+          <div class="text-body-2 mb-2">Type the following to confirm:</div>
+          <div class="fc-token mb-3">{{ expectedToken }}</div>
+          <v-text-field
+            v-model="typed"
+            variant="outlined" density="compact" hide-details
+            autofocus
+            placeholder="paste the token above"
+          />
+        </template>
+      </v-card-text>
+      <v-card-actions>
+        <v-spacer />
+        <v-btn variant="text" @click="onCancel">Cancel</v-btn>
+        <v-btn
+          :color="confirmColor"
+          variant="flat" rounded="pill"
+          :disabled="!canConfirm"
+          @click="onConfirm"
+        >
+          {{ titleVerb }}
+        </v-btn>
+      </v-card-actions>
+    </v-card>
+  </v-dialog>
+</template>
+
+<script setup>
+import { computed, ref, watch } from 'vue'
+
+const props = defineProps({
+  modelValue: { type: Boolean, required: true },
+  action: { type: String, required: true },          // 'restore' | 'delete'
+  kind: { type: String, required: true },            // 'db' | 'images' | 'artist' | 'tag' | 'images-selection'
+  runId: { type: [Number, String], default: '' },    // numeric id or sha8 string
+  description: { type: String, default: '' },
+  tier: { type: String, default: 'C' },              // 'B' | 'C'
+  projectedCounts: { type: Object, default: null },
+})
+const emit = defineEmits(['update:modelValue', 'confirm'])
+
+const typed = ref('')
+const expectedToken = computed(
+  () => `${props.action}-${props.kind}-${props.runId}`,
+)
+const titleVerb = computed(
+  () => props.action === 'restore' ? 'Restore' : 'Delete',
+)
+const kindLabel = computed(() => ({
+  db: 'database backup',
+  images: 'images backup',
+  artist: 'artist',
+  tag: 'tag',
+  'images-selection': 'image selection',
+}[props.kind] || props.kind))
+const alertType = computed(
+  () => props.action === 'restore' ? 'warning' : 'error',
+)
+const confirmColor = computed(
+  () => props.action === 'restore' ? 'warning' : 'error',
+)
+const warningText = computed(() => (
+  props.action === 'restore'
+    ? 'This replaces current state with the backup. There is no undo.'
+    : 'This permanently deletes the listed items. Cannot be recovered.'
+))
+const canConfirm = computed(
+  () => props.tier === 'B'
+    ? true
+    : typed.value === expectedToken.value,
+)
+
+watch(() => props.modelValue, (open) => {
+  if (open) typed.value = ''
+})
+
+function onCancel() {
+  emit('update:modelValue', false)
+}
+function onConfirm() {
+  emit('confirm', expectedToken.value)
+  emit('update:modelValue', false)
+}
+</script>
+
+<style scoped>
+.fc-token {
+  font-family: 'JetBrains Mono', monospace;
+  background: rgb(var(--v-theme-surface-light));
+  padding: 6px 10px; border-radius: 4px;
+  font-size: 14px; word-break: break-all;
+}
+.fc-counts {
+  display: grid; grid-template-columns: max-content auto;
+  gap: 4px 12px;
+  font-size: 13px;
+  color: rgb(var(--v-theme-on-surface-variant));
+}
+.fc-counts-key { font-weight: 500; text-transform: capitalize; }
+.fc-counts-val { font-variant-numeric: tabular-nums; }
+</style>

frontend/src/components/modal/ImageViewer.vue

@@ -2,9 +2,6 @@
   <Teleport to="body">
     <div
       class="fc-viewer" role="dialog" aria-modal="true"
-      @keydown.esc="$emit('close')"
-      @keydown.left="onArrowLeft"
-      @keydown.right="onArrowRight"
       tabindex="-1" ref="rootEl"
     >
       <button class="fc-viewer__close" @click="$emit('close')" aria-label="Close">
@@ -68,7 +65,7 @@ import VideoCanvas from './VideoCanvas.vue'
 import TagPanel from './TagPanel.vue'
 import ProvenancePanel from './ProvenancePanel.vue'
 
-defineEmits(['close'])
+const emit = defineEmits(['close'])
 
 const modal = useModalStore()
 const rootEl = ref(null)
@@ -86,14 +83,39 @@ const integrityBadge = computed(() => {
 })
 
 let prevBodyOverflow = null
+
+// Document-level keyboard handler. Bound on document (not on the modal
+// root) because <video> elements capture focus when the user interacts
+// with their controls — when video has focus, Escape and arrow keys
+// don't bubble reliably to ancestors. Listening on document side-steps
+// that. Filtered via isTextEntry so tag/comment inputs still get their
+// own keystrokes.
+function onKeyDown(ev) {
+  if (ev.key === 'Escape') {
+    if (isTextEntry(ev.target)) return
+    ev.preventDefault()
+    emit('close')
+  } else if (ev.key === 'ArrowLeft') {
+    if (isTextEntry(ev.target)) return
+    ev.preventDefault()
+    modal.goPrev()
+  } else if (ev.key === 'ArrowRight') {
+    if (isTextEntry(ev.target)) return
+    ev.preventDefault()
+    modal.goNext()
+  }
+}
+
 onMounted(async () => {
   prevBodyOverflow = document.body.style.overflow
   document.body.style.overflow = 'hidden'
+  document.addEventListener('keydown', onKeyDown, true)
   await nextFrame()
   rootEl.value?.focus()
 })
 onUnmounted(() => {
   document.body.style.overflow = prevBodyOverflow
+  document.removeEventListener('keydown', onKeyDown, true)
 })
 
 watch(() => modal.currentImageId, async () => {
@@ -105,16 +127,6 @@ function nextFrame() {
   return new Promise(resolve => requestAnimationFrame(resolve))
 }
 
-function onArrowLeft(ev) {
-  if (isTextEntry(ev.target)) return
-  ev.preventDefault()
-  modal.goPrev()
-}
-function onArrowRight(ev) {
-  if (isTextEntry(ev.target)) return
-  ev.preventDefault()
-  modal.goNext()
-}
 function isTextEntry(el) {
   if (!el) return false
   const tag = el.tagName

frontend/src/components/settings/BackupCard.vue

@@ -0,0 +1,205 @@
+<template>
+  <v-card class="fc-backup-card">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-database-export" size="small" />
+      <span>Backups</span>
+    </v-card-title>
+    <v-card-text>
+      <p class="fc-muted text-body-2 mb-4">
+        pg_dump for the database (fast, nightly-schedulable);
+        tar+zstd for the images (slow, manual only). Files live in
+        <code>/images/_backups/</code>. Tag a backup to protect it
+        from autoprune.
+      </p>
+
+      <v-alert
+        v-if="store.lastError"
+        type="warning" variant="tonal" density="compact" class="mb-3"
+      >
+        {{ store.lastError }}
+      </v-alert>
+
+      <!-- Database section -->
+      <h3 class="fc-section-title">Database</h3>
+      <div class="fc-settings-row mb-3">
+        <v-switch
+          :model-value="settings.backup_db_nightly_enabled"
+          color="accent" density="compact" hide-details
+          label="Nightly"
+          @update:model-value="onSettingChange('backup_db_nightly_enabled', $event)"
+        />
+        <v-text-field
+          :model-value="settings.backup_db_nightly_hour_utc"
+          label="Hour (UTC)" type="number" min="0" max="23"
+          density="compact" hide-details style="max-width: 110px;"
+          @update:model-value="onSettingChange('backup_db_nightly_hour_utc', Number($event))"
+        />
+        <v-text-field
+          :model-value="settings.backup_db_keep_last_n"
+          label="Keep last" type="number" min="1" max="365"
+          density="compact" hide-details style="max-width: 110px;"
+          @update:model-value="onSettingChange('backup_db_keep_last_n', Number($event))"
+        />
+      </div>
+      <v-btn
+        color="accent" variant="flat" rounded="pill"
+        prepend-icon="mdi-database-arrow-up" class="mb-3"
+        :loading="dbTriggering"
+        @click="onTrigger('db')"
+      >Run DB backup now</v-btn>
+
+      <BackupRunsTable
+        :runs="store.dbRuns"
+        @restore="onRestore"
+        @delete="onDelete"
+        @tag="onTag"
+      />
+
+      <v-divider class="my-4" />
+
+      <!-- Images section -->
+      <h3 class="fc-section-title">Images</h3>
+      <div class="fc-settings-row mb-3">
+        <v-text-field
+          :model-value="settings.backup_images_keep_last_n"
+          label="Keep last" type="number" min="1" max="100"
+          density="compact" hide-details style="max-width: 110px;"
+          @update:model-value="onSettingChange('backup_images_keep_last_n', Number($event))"
+        />
+      </div>
+      <v-btn
+        color="accent" variant="flat" rounded="pill"
+        prepend-icon="mdi-folder-zip" class="mb-3"
+        :loading="imagesTriggering"
+        @click="onTrigger('images')"
+      >Run images backup now</v-btn>
+      <span class="text-caption fc-muted ml-2">
+        ⓘ Hours for a 200GB+ library.
+      </span>
+
+      <BackupRunsTable
+        :runs="store.imagesRuns"
+        @restore="onRestore"
+        @delete="onDelete"
+        @tag="onTag"
+      />
+    </v-card-text>
+
+    <BackupConfirmModal
+      v-model="confirmOpen"
+      :action="confirmAction"
+      :kind="confirmKind"
+      :run-id="confirmRunId"
+      :description="confirmDescription"
+      @confirm="onConfirmSubmit"
+    />
+  </v-card>
+</template>
+
+<script setup>
+import { computed, onMounted, onUnmounted, ref } from 'vue'
+
+import { useBackupStore } from '../../stores/backup.js'
+import BackupConfirmModal from '../modal/DestructiveConfirmModal.vue'
+import BackupRunsTable from './BackupRunsTable.vue'
+
+const store = useBackupStore()
+
+const settings = computed(() => store.settings || {
+  backup_db_nightly_enabled: false,
+  backup_db_nightly_hour_utc: 3,
+  backup_db_keep_last_n: 14,
+  backup_images_keep_last_n: 3,
+})
+
+const dbTriggering = ref(false)
+const imagesTriggering = ref(false)
+const confirmOpen = ref(false)
+const confirmAction = ref('restore')
+const confirmKind = ref('db')
+const confirmRunId = ref(0)
+const confirmDescription = ref('')
+
+let pollId = null
+function pollOnce() {
+  if (document.hidden) return
+  store.loadRuns('db')
+  store.loadRuns('images')
+}
+
+onMounted(() => {
+  store.loadSettings()
+  pollOnce()
+  pollId = setInterval(pollOnce, 5000)
+})
+onUnmounted(() => {
+  if (pollId) { clearInterval(pollId); pollId = null }
+})
+
+async function onTrigger(kind) {
+  const triggeringRef = kind === 'db' ? dbTriggering : imagesTriggering
+  triggeringRef.value = true
+  try {
+    await store.triggerBackup(kind)
+    pollOnce()
+  } finally {
+    triggeringRef.value = false
+  }
+}
+
+async function onSettingChange(field, value) {
+  await store.patchSettings({ [field]: value })
+}
+
+function _openConfirm(action, run) {
+  confirmAction.value = action
+  confirmKind.value = run.kind
+  confirmRunId.value = run.id
+  confirmDescription.value = (
+    action === 'restore'
+      ? `Source artifact: ${run.kind === 'db' ? run.sql_path : run.tar_path}`
+      : 'Removes the BackupRun row and unlinks the artifact files.'
+  )
+  confirmOpen.value = true
+}
+
+function onRestore(run) { _openConfirm('restore', run) }
+function onDelete(run) { _openConfirm('delete', run) }
+
+async function onTag(run) {
+  const next = prompt(
+    `Tag for backup #${run.id} (blank to clear):`,
+    run.tag || '',
+  )
+  if (next === null) return
+  await store.setTag(run.id, next.trim() || null)
+  pollOnce()
+}
+
+async function onConfirmSubmit(token) {
+  try {
+    if (confirmAction.value === 'restore') {
+      await store.restore(confirmRunId.value, token)
+    } else {
+      await store.deleteRun(confirmRunId.value, token)
+    }
+    pollOnce()
+  } catch {
+    // store.lastError already updated; surfaced in v-alert above.
+  }
+}
+</script>
+
+<style scoped>
+.fc-backup-card { border-radius: 8px; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+.fc-section-title {
+  font-family: 'Fraunces', Georgia, serif;
+  font-size: 16px; font-weight: 500;
+  color: rgb(var(--v-theme-on-surface));
+  margin-bottom: 8px;
+}
+.fc-settings-row {
+  display: flex; gap: 12px; align-items: center; flex-wrap: wrap;
+}
+</style>

frontend/src/components/settings/BackupRunsTable.vue

@@ -0,0 +1,109 @@
+<template>
+  <v-table density="compact" class="fc-backup-table">
+    <thead>
+      <tr>
+        <th>When</th>
+        <th>Status</th>
+        <th class="text-right">Duration</th>
+        <th class="text-right">Size</th>
+        <th>Tag</th>
+        <th class="text-right">Actions</th>
+      </tr>
+    </thead>
+    <tbody>
+      <tr v-for="r in runs" :key="r.id">
+        <td class="fc-tabular" :title="r.started_at">
+          {{ formatRelative(r.started_at) }}
+        </td>
+        <td>
+          <v-icon size="small" :color="statusColor(r.status)">
+            {{ statusIcon(r.status) }}
+          </v-icon>
+          {{ r.status }}
+        </td>
+        <td class="text-right fc-tabular">{{ formatDuration(r) }}</td>
+        <td class="text-right fc-tabular">{{ formatBytes(r.size_bytes) }}</td>
+        <td>
+          <v-chip
+            v-if="r.tag" size="x-small" color="accent" variant="tonal"
+          >{{ r.tag }}</v-chip>
+          <span v-else class="fc-muted">—</span>
+        </td>
+        <td class="text-right">
+          <v-menu>
+            <template #activator="{ props: act }">
+              <v-btn icon="mdi-dots-vertical" size="x-small" variant="text" v-bind="act" />
+            </template>
+            <v-list density="compact">
+              <v-list-item
+                :title="r.tag ? 'Untag' : 'Tag…'"
+                @click="$emit('tag', r)"
+              />
+              <v-list-item
+                title="Restore…" :disabled="r.status !== 'ok'"
+                @click="$emit('restore', r)"
+              />
+              <v-list-item
+                title="Delete…"
+                @click="$emit('delete', r)"
+              />
+            </v-list>
+          </v-menu>
+        </td>
+      </tr>
+      <tr v-if="!runs.length">
+        <td colspan="6" class="text-center fc-muted py-4">No backups yet.</td>
+      </tr>
+    </tbody>
+  </v-table>
+</template>
+
+<script setup>
+defineProps({ runs: { type: Array, default: () => [] } })
+defineEmits(['restore', 'delete', 'tag'])
+
+function statusIcon(s) {
+  return {
+    ok: 'mdi-check-circle', error: 'mdi-close-circle',
+    running: 'mdi-timer-sand', restoring: 'mdi-restore-clock',
+    restored: 'mdi-restore', pending: 'mdi-clock-outline',
+  }[s] || 'mdi-help-circle'
+}
+function statusColor(s) {
+  return {
+    ok: 'success', error: 'error',
+    running: 'accent', restoring: 'info',
+    restored: 'info', pending: 'on-surface-variant',
+  }[s] || 'on-surface-variant'
+}
+function formatDuration(r) {
+  if (r.duration_seconds == null) return '—'
+  const s = r.duration_seconds
+  if (s < 60) return `${s}s`
+  if (s < 3600) return `${Math.floor(s / 60)}m ${s % 60}s`
+  const h = Math.floor(s / 3600); const m = Math.floor((s % 3600) / 60)
+  return `${h}h ${m}m`
+}
+function formatBytes(b) {
+  if (b == null) return '—'
+  const units = ['B', 'KiB', 'MiB', 'GiB', 'TiB']
+  let i = 0; let v = b
+  while (v >= 1024 && i < units.length - 1) { v /= 1024; i++ }
+  return `${v.toFixed(i === 0 ? 0 : 1)} ${units[i]}`
+}
+function formatRelative(iso) {
+  if (!iso) return '—'
+  const then = new Date(iso).getTime()
+  const diff = Math.max(0, (Date.now() - then) / 1000)
+  if (diff < 60) return `${Math.floor(diff)}s ago`
+  if (diff < 3600) return `${Math.floor(diff / 60)}m ago`
+  if (diff < 86400) return `${Math.floor(diff / 3600)}h ago`
+  return `${Math.floor(diff / 86400)}d ago`
+}
+</script>
+
+<style scoped>
+.fc-backup-table { background: transparent; }
+.fc-tabular { font-variant-numeric: tabular-nums; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/components/settings/MaintenancePanel.vue

@@ -12,6 +12,8 @@
     <MLThresholdSliders class="mt-4" />
     <AllowlistTable class="mt-4" />
     <AliasTable class="mt-4" />
+    <BackupCard class="mt-6" />
+    <TagMaintenanceCard class="mt-6" />
     <BrowserExtensionCard class="mt-6" />
     <LegacyMigrationCard class="mt-6" />
   </div>
@@ -23,6 +25,8 @@ import CentroidRecomputeCard from './CentroidRecomputeCard.vue'
 import MLThresholdSliders from './MLThresholdSliders.vue'
 import AllowlistTable from './AllowlistTable.vue'
 import AliasTable from './AliasTable.vue'
+import BackupCard from './BackupCard.vue'
+import TagMaintenanceCard from './TagMaintenanceCard.vue'
 import BrowserExtensionCard from './BrowserExtensionCard.vue'
 import LegacyMigrationCard from './LegacyMigrationCard.vue'
 </script>

frontend/src/components/settings/QueuesTable.vue

@@ -0,0 +1,109 @@
+<template>
+  <v-table density="compact" class="fc-queues-table">
+    <thead>
+      <tr>
+        <th>Queue</th>
+        <th class="text-right">Depth</th>
+        <th class="text-right">Workers</th>
+        <th v-if="!compact" class="text-right">Active</th>
+        <th class="text-right">Last min: ok / err</th>
+      </tr>
+    </thead>
+    <tbody>
+      <tr v-for="q in QUEUE_NAMES" :key="q">
+        <td>{{ q }}</td>
+        <td class="text-right fc-tabular" :class="depthClass(q)">
+          {{ formatDepth(q) }}
+        </td>
+        <td class="text-right fc-tabular">{{ workerCount(q) }}</td>
+        <td v-if="!compact" class="text-right fc-tabular">{{ activeCount(q) }}</td>
+        <td class="text-right fc-tabular">
+          <span class="fc-ok">{{ recentOk(q) }}</span>
+          <span class="fc-sep">/</span>
+          <span :class="recentErr(q) > 0 ? 'fc-err' : 'fc-muted'">
+            {{ recentErr(q) }}<span v-if="recentErr(q) > 0"> ⚠</span>
+          </span>
+        </td>
+      </tr>
+    </tbody>
+  </v-table>
+</template>
+
+<script setup>
+import { computed } from 'vue'
+
+const props = defineProps({
+  queues: { type: Object, default: null },     // store.queues
+  workers: { type: Object, default: null },    // store.workers
+  recentMinute: { type: Array, default: () => [] }, // store.recentMinute
+  compact: { type: Boolean, default: false },
+})
+
+const QUEUE_NAMES = [
+  'default', 'import', 'thumbnail', 'ml',
+  'download', 'scan', 'maintenance',
+]
+
+function formatDepth(name) {
+  const d = props.queues?.queues?.[name]
+  if (d == null) return '—'
+  return d.toLocaleString()
+}
+
+function depthClass(name) {
+  const d = props.queues?.queues?.[name]
+  const workers = workerCount(name)
+  // Queue has depth but no workers → warning.
+  if (d != null && d > 0 && workers === 0) return 'fc-warn'
+  return ''
+}
+
+function workerCount(name) {
+  if (!props.workers?.workers) return 0
+  let count = 0
+  for (const info of Object.values(props.workers.workers)) {
+    if (info.queues?.includes(name)) count++
+  }
+  return count
+}
+
+function activeCount(name) {
+  if (!props.workers?.workers) return 0
+  // Active count is per-worker, not per-queue. Approximate: sum of
+  // active_count across workers that subscribe to this queue.
+  let count = 0
+  for (const info of Object.values(props.workers.workers)) {
+    if (info.queues?.includes(name)) count += info.active_count || 0
+  }
+  return count
+}
+
+const recentByQueue = computed(() => {
+  const out = {}
+  for (const r of props.recentMinute) {
+    if (!out[r.queue]) out[r.queue] = { ok: 0, err: 0 }
+    if (r.status === 'ok') out[r.queue].ok++
+    else if (r.status === 'error' || r.status === 'timeout') out[r.queue].err++
+  }
+  return out
+})
+
+function recentOk(name) { return recentByQueue.value[name]?.ok || 0 }
+function recentErr(name) { return recentByQueue.value[name]?.err || 0 }
+</script>
+
+<style scoped>
+.fc-queues-table { background: transparent; }
+.fc-tabular {
+  font-variant-numeric: tabular-nums;
+  font-feature-settings: 'tnum';
+}
+.fc-ok { color: rgb(var(--v-theme-success, 76 175 80)); }
+.fc-err { color: rgb(var(--v-theme-error, 220 80 80)); }
+.fc-warn { color: rgb(var(--v-theme-warning, 255 179 0)); font-weight: 500; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+.fc-sep {
+  margin: 0 4px;
+  color: rgb(var(--v-theme-on-surface-variant));
+}
+</style>

frontend/src/components/settings/SystemActivitySummary.vue

@@ -0,0 +1,64 @@
+<template>
+  <v-card class="fc-activity-summary">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-pulse" size="small" />
+      <span>System activity (last min)</span>
+      <v-spacer />
+      <v-btn
+        variant="text" size="small" rounded="pill"
+        @click="$emit('open-activity')"
+      >
+        View activity
+        <v-icon end size="small">mdi-arrow-right</v-icon>
+      </v-btn>
+    </v-card-title>
+
+    <v-card-text>
+      <v-alert
+        v-if="store.lastError"
+        type="warning" variant="tonal" density="compact" class="mb-3"
+      >
+        {{ store.lastError }}
+      </v-alert>
+
+      <QueuesTable
+        :queues="store.queues"
+        :workers="store.workers"
+        :recent-minute="store.recentMinute"
+        compact
+      />
+    </v-card-text>
+  </v-card>
+</template>
+
+<script setup>
+import { onMounted, onUnmounted } from 'vue'
+
+import { useSystemActivityStore } from '../../stores/systemActivity.js'
+import QueuesTable from './QueuesTable.vue'
+
+defineEmits(['open-activity'])
+
+const store = useSystemActivityStore()
+
+let pollId = null
+
+function pollOnce() {
+  if (document.hidden) return
+  store.loadQueues()
+  store.loadWorkers()
+  store.loadRecentMinute()
+}
+
+onMounted(() => {
+  pollOnce()
+  pollId = setInterval(pollOnce, 5000)
+})
+onUnmounted(() => {
+  if (pollId) { clearInterval(pollId); pollId = null }
+})
+</script>
+
+<style scoped>
+.fc-activity-summary { border-radius: 8px; }
+</style>

frontend/src/components/settings/SystemActivityTab.vue

@@ -0,0 +1,280 @@
+<template>
+  <div class="fc-activity">
+    <!-- Queues + workers pane -->
+    <v-card class="mb-4">
+      <v-card-title class="d-flex align-center" style="gap: 10px;">
+        <v-icon icon="mdi-pulse" size="small" />
+        <span>Queues + workers</span>
+        <v-spacer />
+        <span class="text-caption fc-muted">
+          updated {{ formatRelative(store.queues?.fetched_at) }}
+        </span>
+      </v-card-title>
+      <v-card-text>
+        <QueuesTable
+          :queues="store.queues"
+          :workers="store.workers"
+          :recent-minute="store.recentMinute"
+        />
+      </v-card-text>
+    </v-card>
+
+    <!-- Recent failures pane -->
+    <v-card class="mb-4">
+      <v-card-title class="d-flex align-center" style="gap: 10px;">
+        <v-icon icon="mdi-alert-circle-outline" size="small" />
+        <span>Recent failures (last 24h)</span>
+        <v-spacer />
+        <span class="text-caption fc-muted">
+          {{ failureCount }} failures
+        </span>
+      </v-card-title>
+      <v-card-text>
+        <div v-if="errorTypes.length" class="mb-3 fc-pills">
+          <v-chip
+            v-for="t in errorTypes" :key="t.name"
+            :color="t.name === filterErrorType ? 'accent' : undefined"
+            size="small" variant="tonal" closable
+            @click="toggleErrorTypeFilter(t.name)"
+            @click:close="toggleErrorTypeFilter(t.name)"
+          >
+            {{ t.name }} × {{ t.count }}
+          </v-chip>
+        </div>
+        <v-table density="compact">
+          <thead>
+            <tr>
+              <th>Time</th>
+              <th>Queue</th>
+              <th>Task</th>
+              <th>Target</th>
+              <th>Error</th>
+            </tr>
+          </thead>
+          <tbody>
+            <tr v-for="r in filteredFailures" :key="r.id">
+              <td class="fc-tabular" :title="r.finished_at">
+                {{ formatRelative(r.finished_at) }}
+              </td>
+              <td>{{ r.queue }}</td>
+              <td><code>{{ shortTaskName(r.task_name) }}</code></td>
+              <td class="fc-tabular">{{ r.target_id ?? '—' }}</td>
+              <td class="fc-err" :title="r.error_message">
+                {{ r.error_type }}
+              </td>
+            </tr>
+            <tr v-if="!filteredFailures.length">
+              <td colspan="5" class="text-center fc-muted py-4">
+                No failures in the last 24h.
+              </td>
+            </tr>
+          </tbody>
+        </v-table>
+      </v-card-text>
+    </v-card>
+
+    <!-- All activity pane -->
+    <v-card>
+      <v-card-title class="d-flex align-center" style="gap: 10px;">
+        <v-icon icon="mdi-format-list-bulleted" size="small" />
+        <span>All recent activity</span>
+        <v-spacer />
+        <v-select
+          v-model="filterQueue"
+          :items="queueOptions" density="compact" hide-details
+          style="max-width: 180px;"
+          @update:model-value="onFilterChange"
+        />
+        <v-select
+          v-model="filterStatus"
+          :items="statusOptions" density="compact" hide-details
+          style="max-width: 180px;"
+          @update:model-value="onFilterChange"
+        />
+        <v-btn
+          variant="text" size="small" rounded="pill"
+          @click="onRefresh"
+        >
+          <v-icon start size="small">mdi-refresh</v-icon>
+          Refresh
+        </v-btn>
+      </v-card-title>
+      <v-card-text>
+        <v-table density="compact">
+          <thead>
+            <tr>
+              <th></th>
+              <th>Time</th>
+              <th>Queue</th>
+              <th>Task</th>
+              <th>Target</th>
+              <th class="text-right">Duration</th>
+              <th>Status</th>
+            </tr>
+          </thead>
+          <tbody>
+            <tr v-for="r in store.runs" :key="r.id">
+              <td><v-icon size="small" :color="statusColor(r.status)">{{ statusIcon(r.status) }}</v-icon></td>
+              <td class="fc-tabular" :title="r.started_at">{{ formatRelative(r.started_at) }}</td>
+              <td>{{ r.queue }}</td>
+              <td><code>{{ shortTaskName(r.task_name) }}</code></td>
+              <td class="fc-tabular">{{ r.target_id ?? '—' }}</td>
+              <td class="text-right fc-tabular">{{ formatDuration(r.duration_ms) }}</td>
+              <td>{{ r.status }}</td>
+            </tr>
+            <tr v-if="!store.runs.length && !store.loading.runs">
+              <td colspan="7" class="text-center fc-muted py-4">
+                No activity yet.
+              </td>
+            </tr>
+          </tbody>
+        </v-table>
+        <div v-if="store.runsHasMore" class="d-flex justify-center py-3">
+          <v-btn
+            variant="text" size="small"
+            :loading="store.loading.runs"
+            @click="onLoadMore"
+          >Load more</v-btn>
+        </div>
+      </v-card-text>
+    </v-card>
+  </div>
+</template>
+
+<script setup>
+import { computed, onMounted, onUnmounted, ref } from 'vue'
+
+import { useSystemActivityStore } from '../../stores/systemActivity.js'
+import QueuesTable from './QueuesTable.vue'
+
+const store = useSystemActivityStore()
+
+const filterQueue = ref(null)
+const filterStatus = ref(null)
+const filterErrorType = ref(null)
+
+const queueOptions = [
+  { title: 'All queues', value: null },
+  { title: 'import', value: 'import' },
+  { title: 'thumbnail', value: 'thumbnail' },
+  { title: 'ml', value: 'ml' },
+  { title: 'download', value: 'download' },
+  { title: 'scan', value: 'scan' },
+  { title: 'maintenance', value: 'maintenance' },
+  { title: 'default', value: 'default' },
+]
+const statusOptions = [
+  { title: 'All statuses', value: null },
+  { title: 'Running', value: 'running' },
+  { title: 'OK', value: 'ok' },
+  { title: 'Error', value: 'error' },
+  { title: 'Timeout', value: 'timeout' },
+  { title: 'Retry', value: 'retry' },
+]
+
+let queuesPollId = null
+let failuresPollId = null
+
+function pollQueues() {
+  if (document.hidden) return
+  store.loadQueues()
+  store.loadWorkers()
+  store.loadRecentMinute()
+}
+function pollFailures() {
+  if (document.hidden) return
+  store.loadFailures()
+}
+
+onMounted(() => {
+  pollQueues()
+  pollFailures()
+  store.setFilter({ queue: null, status: null })
+  store.loadRuns({ reset: true })
+  queuesPollId = setInterval(pollQueues, 3000)
+  failuresPollId = setInterval(pollFailures, 15000)
+})
+onUnmounted(() => {
+  if (queuesPollId) clearInterval(queuesPollId)
+  if (failuresPollId) clearInterval(failuresPollId)
+})
+
+const errorTypes = computed(() => {
+  const counts = store.failures?.count_by_type || {}
+  return Object.entries(counts).map(([name, count]) => ({ name, count }))
+})
+
+const failureCount = computed(() => (store.failures?.recent || []).length)
+
+const filteredFailures = computed(() => {
+  const all = store.failures?.recent || []
+  if (!filterErrorType.value) return all
+  return all.filter(r => r.error_type === filterErrorType.value)
+})
+
+function toggleErrorTypeFilter(name) {
+  filterErrorType.value = filterErrorType.value === name ? null : name
+  if (filterErrorType.value) {
+    // Mirror into the All-activity table too.
+    filterStatus.value = 'error'
+    onFilterChange()
+  }
+}
+
+function onFilterChange() {
+  store.setFilter({ queue: filterQueue.value, status: filterStatus.value })
+  store.loadRuns({ reset: true })
+}
+function onLoadMore() { store.loadRuns({ reset: false }) }
+function onRefresh() { store.loadRuns({ reset: true }) }
+
+function statusIcon(status) {
+  return {
+    ok: 'mdi-check-circle',
+    error: 'mdi-close-circle',
+    timeout: 'mdi-clock-alert',
+    retry: 'mdi-refresh',
+    running: 'mdi-timer-sand',
+  }[status] || 'mdi-help-circle'
+}
+function statusColor(status) {
+  return {
+    ok: 'success', error: 'error', timeout: 'warning',
+    retry: 'info', running: 'accent',
+  }[status] || 'on-surface-variant'
+}
+function shortTaskName(name) {
+  if (!name) return ''
+  const parts = name.split('.')
+  return parts.slice(-1)[0]
+}
+function formatDuration(ms) {
+  if (ms == null) return '—'
+  if (ms < 1000) return `${ms} ms`
+  if (ms < 60_000) return `${(ms / 1000).toFixed(1)} s`
+  return `${(ms / 60_000).toFixed(1)} min`
+}
+function formatRelative(iso) {
+  if (!iso) return '—'
+  const then = new Date(iso).getTime()
+  const now = Date.now()
+  const diff = Math.max(0, (now - then) / 1000)
+  if (diff < 60) return `${Math.floor(diff)}s ago`
+  if (diff < 3600) return `${Math.floor(diff / 60)}m ago`
+  if (diff < 86400) return `${Math.floor(diff / 3600)}h ago`
+  return `${Math.floor(diff / 86400)}d ago`
+}
+</script>
+
+<style scoped>
+.fc-activity { padding: 0; }
+.fc-pills {
+  display: flex; gap: 6px; flex-wrap: wrap;
+}
+.fc-tabular {
+  font-variant-numeric: tabular-nums;
+  font-feature-settings: 'tnum';
+}
+.fc-err { color: rgb(var(--v-theme-error, 220 80 80)); }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/components/settings/TagMaintenanceCard.vue

@@ -0,0 +1,95 @@
+<template>
+  <v-card class="fc-tag-maint">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-tag-remove" size="small" />
+      <span>Tag maintenance</span>
+    </v-card-title>
+    <v-card-text>
+      <p class="fc-muted text-body-2 mb-4">
+        Remove tags with zero image associations and zero series-page
+        references. Auto-created tag rows that never got applied get
+        swept here.
+      </p>
+
+      <v-alert
+        v-if="store.lastError"
+        type="warning" variant="tonal" density="compact" class="mb-3"
+      >{{ store.lastError }}</v-alert>
+
+      <v-btn
+        color="accent" variant="flat" rounded="pill"
+        prepend-icon="mdi-magnify"
+        :loading="loadingPreview"
+        class="mb-3"
+        @click="onPreview"
+      >Preview unused tags</v-btn>
+
+      <div v-if="preview">
+        <p class="text-body-2 mb-2">
+          <strong>{{ preview.count }}</strong> unused tag(s).
+          <span v-if="preview.count > 50" class="fc-muted">
+            Showing first 50 names.
+          </span>
+        </p>
+        <div v-if="preview.sample_names?.length" class="fc-name-grid mb-3">
+          <span v-for="n in preview.sample_names" :key="n" class="fc-name">
+            {{ n }}
+          </span>
+        </div>
+        <v-btn
+          color="error" variant="flat" rounded="pill"
+          prepend-icon="mdi-delete-sweep"
+          :disabled="!preview.count"
+          :loading="committing"
+          @click="onCommit"
+        >Delete {{ preview.count }} unused tag(s)</v-btn>
+      </div>
+    </v-card-text>
+  </v-card>
+</template>
+
+<script setup>
+import { ref } from 'vue'
+
+import { useAdminStore } from '../../stores/admin.js'
+
+const store = useAdminStore()
+const preview = ref(null)
+const loadingPreview = ref(false)
+const committing = ref(false)
+
+async function onPreview() {
+  loadingPreview.value = true
+  try {
+    preview.value = await store.pruneUnusedTags({ dryRun: true })
+  } finally {
+    loadingPreview.value = false
+  }
+}
+
+async function onCommit() {
+  committing.value = true
+  try {
+    const result = await store.pruneUnusedTags({ dryRun: false })
+    preview.value = { count: 0, sample_names: result.sample_names || [] }
+  } finally {
+    committing.value = false
+  }
+}
+</script>
+
+<style scoped>
+.fc-tag-maint { border-radius: 8px; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+.fc-name-grid {
+  display: flex; flex-wrap: wrap; gap: 4px 8px;
+  max-height: 200px; overflow-y: auto;
+  padding: 8px; border-radius: 4px;
+  background: rgb(var(--v-theme-surface-light));
+}
+.fc-name {
+  font-family: 'JetBrains Mono', monospace;
+  font-size: 12px;
+  color: rgb(var(--v-theme-on-surface-variant));
+}
+</style>

frontend/src/stores/admin.js

@@ -0,0 +1,153 @@
+import { defineStore } from 'pinia'
+import { ref } from 'vue'
+
+import { useApi } from '../composables/useApi.js'
+
+export const useAdminStore = defineStore('admin', () => {
+  const api = useApi()
+  const lastError = ref(null)
+
+  // --- Tier-C: artist cascade ---------------------------------------
+
+  async function projectArtistCascade(slug) {
+    lastError.value = null
+    try {
+      return await api.post(
+        `/api/admin/artists/${encodeURIComponent(slug)}/cascade-delete`,
+        { body: { dry_run: true } },
+      )
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function dispatchArtistCascade(slug, confirm) {
+    lastError.value = null
+    try {
+      return await api.post(
+        `/api/admin/artists/${encodeURIComponent(slug)}/cascade-delete`,
+        { body: { dry_run: false, confirm } },
+      )
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  // --- Tier-C: bulk image delete ------------------------------------
+
+  async function projectBulkImageDelete(imageIds) {
+    lastError.value = null
+    try {
+      return await api.post(
+        '/api/admin/images/bulk-delete',
+        { body: { image_ids: imageIds, dry_run: true } },
+      )
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function dispatchBulkImageDelete(imageIds, confirm) {
+    lastError.value = null
+    try {
+      return await api.post(
+        '/api/admin/images/bulk-delete',
+        { body: { image_ids: imageIds, dry_run: false, confirm } },
+      )
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  // --- Tier-B: tag delete + merge -----------------------------------
+
+  async function deleteTag(tagId) {
+    lastError.value = null
+    try {
+      return await api.delete(`/api/admin/tags/${tagId}`)
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function mergeTags(destId, sourceId) {
+    lastError.value = null
+    try {
+      return await api.post(
+        `/api/admin/tags/${destId}/merge`,
+        { body: { source_id: sourceId } },
+      )
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function tagUsageCount(tagId) {
+    lastError.value = null
+    try {
+      const body = await api.get(`/api/admin/tags/${tagId}/usage-count`)
+      return body.count
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  // --- Tier-A: prune unused -----------------------------------------
+
+  async function pruneUnusedTags({ dryRun = true } = {}) {
+    lastError.value = null
+    try {
+      return await api.post(
+        '/api/admin/tags/prune-unused',
+        { body: { dry_run: dryRun } },
+      )
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  // --- Task progress polling (taps FC-3i activity dashboard) --------
+
+  /**
+   * Polls /api/system/activity/runs?queue=maintenance every 3s,
+   * resolves when a task_run row with the given celery task_id
+   * reaches a terminal status (ok / error / timeout). Returns the
+   * row. Times out after 30 min by default.
+   */
+  async function pollTaskUntilDone(taskId, { timeoutMs = 1_800_000 } = {}) {
+    const deadline = Date.now() + timeoutMs
+    while (Date.now() < deadline) {
+      const body = await api.get(
+        '/api/system/activity/runs',
+        { params: { queue: 'maintenance', limit: 20 } },
+      )
+      const row = (body.runs || []).find(r => r.celery_task_id === taskId)
+      if (row && ['ok', 'error', 'timeout'].includes(row.status)) {
+        return row
+      }
+      await new Promise((r) => setTimeout(r, 3000))
+    }
+    throw new Error(`task ${taskId} did not finish within ${timeoutMs}ms`)
+  }
+
+  return {
+    lastError,
+    projectArtistCascade,
+    dispatchArtistCascade,
+    projectBulkImageDelete,
+    dispatchBulkImageDelete,
+    deleteTag,
+    mergeTags,
+    tagUsageCount,
+    pruneUnusedTags,
+    pollTaskUntilDone,
+  }
+})

frontend/src/stores/backup.js

@@ -0,0 +1,109 @@
+import { defineStore } from 'pinia'
+import { ref } from 'vue'
+
+import { useApi } from '../composables/useApi.js'
+
+export const useBackupStore = defineStore('backup', () => {
+  const api = useApi()
+
+  const dbRuns = ref([])
+  const imagesRuns = ref([])
+  const settings = ref(null)
+  const loading = ref({ dbRuns: false, imagesRuns: false, settings: false })
+  const lastError = ref(null)
+
+  async function loadRuns(kind) {
+    const targetRef = kind === 'db' ? dbRuns : imagesRuns
+    const loadingKey = kind === 'db' ? 'dbRuns' : 'imagesRuns'
+    loading.value[loadingKey] = true
+    lastError.value = null
+    try {
+      const body = await api.get('/api/system/backup/runs', {
+        params: { kind, limit: 50 },
+      })
+      targetRef.value = body.runs || []
+    } catch (e) {
+      lastError.value = e.message
+    } finally {
+      loading.value[loadingKey] = false
+    }
+  }
+
+  async function triggerBackup(kind, tag = null) {
+    lastError.value = null
+    try {
+      await api.post(`/api/system/backup/${kind}`, {
+        body: tag ? { tag } : {},
+      })
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function restore(runId, confirm) {
+    lastError.value = null
+    try {
+      await api.post(`/api/system/backup/runs/${runId}/restore`, {
+        body: { confirm },
+      })
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function deleteRun(runId, confirm) {
+    lastError.value = null
+    try {
+      await api.delete(`/api/system/backup/runs/${runId}`, {
+        body: { confirm },
+      })
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function setTag(runId, tag) {
+    lastError.value = null
+    try {
+      await api.patch(`/api/system/backup/runs/${runId}`, {
+        body: { tag },
+      })
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  async function loadSettings() {
+    loading.value.settings = true
+    lastError.value = null
+    try {
+      settings.value = await api.get('/api/system/backup/settings')
+    } catch (e) {
+      lastError.value = e.message
+    } finally {
+      loading.value.settings = false
+    }
+  }
+
+  async function patchSettings(patch) {
+    lastError.value = null
+    try {
+      settings.value = await api.patch('/api/system/backup/settings', {
+        body: patch,
+      })
+    } catch (e) {
+      lastError.value = e.message
+      throw e
+    }
+  }
+
+  return {
+    dbRuns, imagesRuns, settings, loading, lastError,
+    loadRuns, triggerBackup, restore, deleteRun, setTag,
+    loadSettings, patchSettings,
+  }
+})

frontend/src/stores/systemActivity.js

@@ -0,0 +1,103 @@
+import { defineStore } from 'pinia'
+import { ref } from 'vue'
+
+import { useApi } from '../composables/useApi.js'
+
+export const useSystemActivityStore = defineStore('systemActivity', () => {
+  const api = useApi()
+
+  // Live polled state.
+  const queues = ref(null)       // { queues: {name: depth|null}, fetched_at }
+  const workers = ref(null)      // { workers: {hostname: {...}}, fetched_at }
+  const recentMinute = ref([])   // last-60s rows (for Overview summary)
+  const failures = ref(null)     // { recent, count_by_type, since }
+
+  // Paginated runs (Activity tab "All recent activity" pane).
+  const runs = ref([])
+  const runsCursor = ref(null)
+  const runsHasMore = ref(false)
+  const runsFilter = ref({ queue: null, status: null, limit: 50 })
+
+  const loading = ref({ queues: false, workers: false, runs: false, failures: false })
+  const lastError = ref(null)
+
+  async function loadQueues() {
+    loading.value.queues = true
+    lastError.value = null
+    try {
+      queues.value = await api.get('/api/system/activity/queues')
+    } catch (e) {
+      lastError.value = e.message
+    } finally {
+      loading.value.queues = false
+    }
+  }
+
+  async function loadWorkers() {
+    loading.value.workers = true
+    lastError.value = null
+    try {
+      workers.value = await api.get('/api/system/activity/workers')
+    } catch (e) {
+      lastError.value = e.message
+    } finally {
+      loading.value.workers = false
+    }
+  }
+
+  async function loadRecentMinute() {
+    // Used by the Overview summary card: pull last 60s of runs to compute
+    // per-queue ok/err counts. One call covers all queues; UI groups.
+    try {
+      const body = await api.get('/api/system/activity/runs', {
+        params: { limit: 200 },
+      })
+      recentMinute.value = body.runs || []
+    } catch (e) {
+      lastError.value = e.message
+    }
+  }
+
+  async function loadRuns({ reset = false } = {}) {
+    loading.value.runs = true
+    lastError.value = null
+    try {
+      const params = { limit: runsFilter.value.limit }
+      if (runsFilter.value.queue) params.queue = runsFilter.value.queue
+      if (runsFilter.value.status) params.status = runsFilter.value.status
+      if (!reset && runsCursor.value) params.before_id = runsCursor.value
+      const body = await api.get('/api/system/activity/runs', { params })
+      runs.value = reset ? body.runs : [...runs.value, ...body.runs]
+      runsCursor.value = body.next_cursor
+      runsHasMore.value = body.next_cursor !== null
+    } catch (e) {
+      lastError.value = e.message
+    } finally {
+      loading.value.runs = false
+    }
+  }
+
+  async function loadFailures() {
+    loading.value.failures = true
+    lastError.value = null
+    try {
+      failures.value = await api.get('/api/system/activity/failures')
+    } catch (e) {
+      lastError.value = e.message
+    } finally {
+      loading.value.failures = false
+    }
+  }
+
+  function setFilter(filter) {
+    runsFilter.value = { ...runsFilter.value, ...filter }
+  }
+
+  return {
+    queues, workers, recentMinute, failures,
+    runs, runsCursor, runsHasMore, runsFilter,
+    loading, lastError,
+    loadQueues, loadWorkers, loadRecentMinute,
+    loadRuns, loadFailures, setFilter,
+  }
+})

frontend/src/views/ArtistView.vue

@@ -96,6 +96,12 @@
           @open="openImage"
         />
       </section>
+
+      <ArtistDangerZone
+        :slug="slug"
+        :artist-id="store.overview.id"
+        :artist-name="store.overview.name"
+      />
     </template>
   </v-container>
 </template>
@@ -106,6 +112,7 @@ import { useRoute, useRouter, RouterLink } from 'vue-router'
 import { useArtistStore } from '../stores/artist.js'
 import { useModalStore } from '../stores/modal.js'
 import MasonryGrid from '../components/discovery/MasonryGrid.vue'
+import ArtistDangerZone from '../components/artist/ArtistDangerZone.vue'
 
 const route = useRoute()
 const router = useRouter()

frontend/src/views/SettingsView.vue

@@ -2,6 +2,7 @@
   <v-container fluid class="py-6">
     <v-tabs v-model="tab" color="accent" class="mb-4">
       <v-tab value="overview">Overview</v-tab>
+      <v-tab value="activity">Activity</v-tab>
       <v-tab value="import">Import</v-tab>
       <v-tab value="maintenance">Maintenance</v-tab>
     </v-tabs>
@@ -9,6 +10,10 @@
     <v-window v-model="tab">
       <v-window-item value="overview">
         <SystemStatsCards :stats="system.stats" />
+        <SystemActivitySummary
+          class="mt-4"
+          @open-activity="tab = 'activity'"
+        />
         <v-alert
           v-if="system.stats && (system.stats.tasks.pending + system.stats.tasks.queued) > 0"
           type="info" variant="tonal" class="mt-4" closable
@@ -18,6 +23,10 @@
         </v-alert>
       </v-window-item>
 
+      <v-window-item value="activity">
+        <SystemActivityTab />
+      </v-window-item>
+
       <v-window-item value="import">
         <ImportTriggerPanel />
         <v-divider class="my-6" />
@@ -38,6 +47,8 @@ import { onMounted, onUnmounted, ref, watch } from 'vue'
 import { useSystemStore } from '../stores/system.js'
 import { useImportStore } from '../stores/import.js'
 import SystemStatsCards from '../components/settings/SystemStatsCards.vue'
+import SystemActivitySummary from '../components/settings/SystemActivitySummary.vue'
+import SystemActivityTab from '../components/settings/SystemActivityTab.vue'
 import ImportTriggerPanel from '../components/settings/ImportTriggerPanel.vue'
 import ImportFiltersForm from '../components/settings/ImportFiltersForm.vue'
 import ImportTaskList from '../components/settings/ImportTaskList.vue'

frontend/src/views/TagsView.vue

@@ -27,6 +27,7 @@
       <TagCard
         v-for="c in store.cards" :key="c.id" :card="c"
         @open="openTag" @rename="onRename" @manage="onManage" @read="onRead"
+        @merge-with="onMergeWith" @delete="onDeleteTag"
       />
     </div>
 
@@ -41,6 +42,49 @@
       @confirm="confirmMerge"
       @cancel="pendingMerge = null"
     />
+
+    <v-dialog v-model="mergePickerOpen" max-width="520">
+      <v-card>
+        <v-card-title>Merge “{{ mergeSource?.name }}” into…</v-card-title>
+        <v-card-text>
+          <p class="text-body-2 mb-3" style="opacity: 0.7;">
+            Pick the target tag. Source tag will be deleted; all its
+            image associations will move to the target. Must be same
+            kind ({{ mergeSource?.kind }}).
+          </p>
+          <v-autocomplete
+            v-model="mergeTargetId"
+            :items="mergeHits" item-title="name" item-value="id"
+            :loading="mergeLoading"
+            density="compact" variant="outlined" hide-details
+            placeholder="Search target tag…" no-filter
+            @update:search="onMergeSearch"
+          />
+        </v-card-text>
+        <v-card-actions>
+          <v-spacer />
+          <v-btn variant="text" @click="mergePickerOpen = false">Cancel</v-btn>
+          <v-btn
+            color="warning" variant="flat" rounded="pill"
+            :disabled="!mergeTargetId"
+            @click="onMergeConfirm"
+          >Merge</v-btn>
+        </v-card-actions>
+      </v-card>
+    </v-dialog>
+
+    <DestructiveConfirmModal
+      v-model="deleteTagModalOpen"
+      action="delete"
+      kind="tag"
+      :run-id="deleteTagTarget?.id || 0"
+      tier="B"
+      :projected-counts="{ associations: deleteTagUsage }"
+      :description="deleteTagTarget
+        ? `Delete tag “${deleteTagTarget.name}” and remove all image associations.`
+        : ''"
+      @confirm="onDeleteTagConfirm"
+    />
   </v-container>
 </template>
 
@@ -48,8 +92,11 @@
 import { ref, watch, onMounted, onUnmounted } from 'vue'
 import { useRouter } from 'vue-router'
 import { useTagDirectoryStore } from '../stores/tagDirectory.js'
+import { useAdminStore } from '../stores/admin.js'
+import { useApi } from '../composables/useApi.js'
 import TagCard from '../components/discovery/TagCard.vue'
 import MergeConfirmDialog from '../components/discovery/MergeConfirmDialog.vue'
+import DestructiveConfirmModal from '../components/modal/DestructiveConfirmModal.vue'
 
 // Must stay a subset of the backend TagKind enum (character, fandom,
 // general, series, archive, post, meta, rating). 'fandom' is this
@@ -106,6 +153,80 @@ function onManage(id) {
 function onRead(id) {
   router.push({ name: 'series-read', params: { tagId: id } })
 }
+
+// --- FC-3k tag merge + delete ----------------------------------------
+
+const adminStore = useAdminStore()
+const api = useApi()
+
+// Tag merge via dots-menu (separate from inline-rename collision flow above)
+const mergePickerOpen = ref(false)
+const mergeSource = ref(null)
+const mergeTargetId = ref(null)
+const mergeHits = ref([])
+const mergeLoading = ref(false)
+let mergeDebounce = null
+
+function onMergeWith(card) {
+  mergeSource.value = card
+  mergeTargetId.value = null
+  mergeHits.value = []
+  mergePickerOpen.value = true
+}
+
+function onMergeSearch(q) {
+  if (mergeDebounce) clearTimeout(mergeDebounce)
+  if (!q) { mergeHits.value = []; return }
+  mergeDebounce = setTimeout(async () => {
+    mergeLoading.value = true
+    try {
+      const hits = await api.get('/api/tags/autocomplete', {
+        params: { q, kind: mergeSource.value?.kind, limit: 20 },
+      })
+      mergeHits.value = (hits || []).filter(
+        (h) => h.id !== mergeSource.value?.id,
+      )
+    } finally {
+      mergeLoading.value = false
+    }
+  }, 250)
+}
+
+async function onMergeConfirm() {
+  if (!mergeSource.value || !mergeTargetId.value) return
+  try {
+    await adminStore.mergeTags(mergeTargetId.value, mergeSource.value.id)
+    mergePickerOpen.value = false
+    store.reset()
+  } catch {
+    // adminStore.lastError already set; UI surfaces it elsewhere.
+  }
+}
+
+// Tag delete via dots-menu (Tier B)
+const deleteTagModalOpen = ref(false)
+const deleteTagTarget = ref(null)
+const deleteTagUsage = ref(0)
+
+async function onDeleteTag(card) {
+  deleteTagTarget.value = card
+  try {
+    deleteTagUsage.value = await adminStore.tagUsageCount(card.id)
+  } catch {
+    deleteTagUsage.value = 0
+  }
+  deleteTagModalOpen.value = true
+}
+
+async function onDeleteTagConfirm() {
+  if (!deleteTagTarget.value) return
+  try {
+    await adminStore.deleteTag(deleteTagTarget.value.id)
+    store.reset()
+  } catch {
+    // adminStore.lastError already set.
+  }
+}
 </script>
 
 <style scoped>

tests/test_api_admin.py

@@ -0,0 +1,361 @@
+"""FC-3k: /api/admin/* endpoint integration tests.
+
+Monkeypatch `.delay()` for Tier-C tasks to capture dispatch payloads
+without actually queuing. Tier-B/A endpoints run synchronously
+through the real service.
+"""
+import hashlib
+
+import pytest
+
+from backend.app import create_app
+from backend.app.models import Artist, ImageRecord, Tag, TagKind
+from backend.app.models.tag import image_tag
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture
+async def app():
+    return create_app()
+
+
+@pytest.fixture
+async def client(app):
+    async with app.test_client() as c:
+        yield c
+
+
+# --- Tier-C: POST /artists/<slug>/cascade-delete --------------------
+
+
+@pytest.mark.asyncio
+async def test_artist_cascade_dry_run_returns_projection(client, db):
+    a = Artist(name="Aria", slug="aria")
+    db.add(a)
+    await db.commit()
+
+    resp = await client.post(
+        "/api/admin/artists/aria/cascade-delete",
+        json={"dry_run": True},
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["artist"]["slug"] == "aria"
+    assert "projected" in body
+    assert body["projected"]["images"] == 0
+
+
+@pytest.mark.asyncio
+async def test_artist_cascade_unknown_slug_404(client):
+    resp = await client.post(
+        "/api/admin/artists/no-such/cascade-delete",
+        json={"dry_run": True},
+    )
+    assert resp.status_code == 404
+
+
+@pytest.mark.asyncio
+async def test_artist_cascade_wrong_confirm_400(client, db):
+    a = Artist(name="Aria", slug="aria")
+    db.add(a)
+    await db.commit()
+    artist_id = a.id
+
+    resp = await client.post(
+        "/api/admin/artists/aria/cascade-delete",
+        json={"dry_run": False, "confirm": "wrong"},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "confirm_mismatch"
+    assert body["expected"] == f"delete-artist-{artist_id}"
+
+
+@pytest.mark.asyncio
+async def test_artist_cascade_correct_confirm_dispatches(
+    client, db, monkeypatch,
+):
+    a = Artist(name="Aria", slug="aria")
+    db.add(a)
+    await db.commit()
+    artist_id = a.id
+
+    dispatched = []
+
+    class _Result:
+        id = "fake-task-id"
+
+    def _delay(**kw):
+        dispatched.append(kw)
+        return _Result()
+
+    monkeypatch.setattr(
+        "backend.app.tasks.admin.delete_artist_cascade_task.delay", _delay,
+    )
+
+    resp = await client.post(
+        "/api/admin/artists/aria/cascade-delete",
+        json={
+            "dry_run": False,
+            "confirm": f"delete-artist-{artist_id}",
+        },
+    )
+    assert resp.status_code == 202
+    body = await resp.get_json()
+    assert body["task_id"] == "fake-task-id"
+    assert dispatched == [{"artist_id": artist_id}]
+
+
+# --- Tier-C: POST /images/bulk-delete -------------------------------
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_dry_run_returns_counts(client, db, tmp_path):
+    a = Artist(name="B", slug="b")
+    db.add(a)
+    await db.flush()
+    i1 = ImageRecord(
+        artist_id=a.id, path=str(tmp_path / "1.jpg"),
+        sha256="1" * 64, size_bytes=10, mime="image/jpeg",
+        origin="imported_filesystem",
+    )
+    i2 = ImageRecord(
+        artist_id=a.id, path=str(tmp_path / "2.jpg"),
+        sha256="2" * 64, size_bytes=20, mime="image/jpeg",
+        origin="imported_filesystem",
+    )
+    db.add_all([i1, i2])
+    await db.commit()
+
+    resp = await client.post(
+        "/api/admin/images/bulk-delete",
+        json={
+            "image_ids": [i1.id, i2.id, 9_999_999],
+            "dry_run": True,
+        },
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["images_found"] == 2
+    assert body["bytes_on_disk"] == 30
+    assert body["missing_ids"] == [9_999_999]
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_empty_list_400(client):
+    resp = await client.post(
+        "/api/admin/images/bulk-delete",
+        json={"image_ids": [], "dry_run": True},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "invalid_image_ids"
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_non_int_400(client):
+    resp = await client.post(
+        "/api/admin/images/bulk-delete",
+        json={"image_ids": ["foo", 2], "dry_run": True},
+    )
+    assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_wrong_confirm_400_with_expected_token(
+    client, db, tmp_path,
+):
+    a = Artist(name="B", slug="b")
+    db.add(a)
+    await db.flush()
+    img = ImageRecord(
+        artist_id=a.id, path=str(tmp_path / "x.jpg"),
+        sha256="c" * 64, size_bytes=10, mime="image/jpeg",
+        origin="imported_filesystem",
+    )
+    db.add(img)
+    await db.commit()
+    img_id = img.id
+
+    resp = await client.post(
+        "/api/admin/images/bulk-delete",
+        json={
+            "image_ids": [img_id], "dry_run": False, "confirm": "nope",
+        },
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "confirm_mismatch"
+    assert body["expected"].startswith("delete-images-")
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_correct_confirm_dispatches(
+    client, db, tmp_path, monkeypatch,
+):
+    a = Artist(name="B", slug="b")
+    db.add(a)
+    await db.flush()
+    img = ImageRecord(
+        artist_id=a.id, path=str(tmp_path / "x.jpg"),
+        sha256="d" * 64, size_bytes=10, mime="image/jpeg",
+        origin="imported_filesystem",
+    )
+    db.add(img)
+    await db.commit()
+    img_id = img.id
+
+    sha8 = hashlib.sha256(str(img_id).encode("utf-8")).hexdigest()[:8]
+
+    dispatched = []
+
+    class _Result:
+        id = "fake-bulk-id"
+
+    def _delay(**kw):
+        dispatched.append(kw)
+        return _Result()
+
+    monkeypatch.setattr(
+        "backend.app.tasks.admin.bulk_delete_images_task.delay", _delay,
+    )
+
+    resp = await client.post(
+        "/api/admin/images/bulk-delete",
+        json={
+            "image_ids": [img_id],
+            "dry_run": False,
+            "confirm": f"delete-images-{sha8}",
+        },
+    )
+    assert resp.status_code == 202
+    body = await resp.get_json()
+    assert body["task_id"] == "fake-bulk-id"
+    assert dispatched == [{"image_ids": [img_id]}]
+
+
+# --- Tier-B: DELETE /tags/<id> + POST /tags/<dest>/merge -----------
+
+
+@pytest.mark.asyncio
+async def test_tag_delete_200_after_cascade(client, db):
+    t = Tag(name="doomed-tag", kind=TagKind.general)
+    db.add(t)
+    await db.commit()
+    tag_id = t.id
+
+    resp = await client.delete(f"/api/admin/tags/{tag_id}")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["deleted"]["id"] == tag_id
+    assert body["associations_removed"] == 0
+
+
+@pytest.mark.asyncio
+async def test_tag_delete_404_unknown(client):
+    resp = await client.delete("/api/admin/tags/9999999")
+    assert resp.status_code == 404
+
+
+@pytest.mark.asyncio
+async def test_tag_merge_succeeds_for_same_kind(client, db):
+    src = Tag(name="src-tag", kind=TagKind.general)
+    dest = Tag(name="dest-tag", kind=TagKind.general)
+    db.add_all([src, dest])
+    await db.commit()
+    src_id = src.id
+    dest_id = dest.id
+
+    resp = await client.post(
+        f"/api/admin/tags/{dest_id}/merge",
+        json={"source_id": src_id},
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["result"]["target_id"] == dest_id
+
+
+@pytest.mark.asyncio
+async def test_tag_merge_rejects_self_merge(client, db):
+    t = Tag(name="x", kind=TagKind.general)
+    db.add(t)
+    await db.commit()
+    resp = await client.post(
+        f"/api/admin/tags/{t.id}/merge",
+        json={"source_id": t.id},
+    )
+    assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+async def test_tag_merge_returns_400_on_kind_mismatch(client, db):
+    src = Tag(name="src", kind=TagKind.general)
+    dest = Tag(name="dest", kind=TagKind.artist)
+    db.add_all([src, dest])
+    await db.commit()
+    resp = await client.post(
+        f"/api/admin/tags/{dest.id}/merge",
+        json={"source_id": src.id},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "tag_kind_mismatch"
+
+
+# --- Tier-B helper: GET /tags/<id>/usage-count ----------------------
+
+
+@pytest.mark.asyncio
+async def test_tag_usage_count_returns_zero_for_unused(client, db):
+    t = Tag(name="lonely", kind=TagKind.general)
+    db.add(t)
+    await db.commit()
+    resp = await client.get(f"/api/admin/tags/{t.id}/usage-count")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["count"] == 0
+
+
+# --- Tier-A: POST /tags/prune-unused --------------------------------
+
+
+@pytest.mark.asyncio
+async def test_prune_unused_dry_run_lists_unused(client, db, tmp_path):
+    a = Artist(name="P", slug="p")
+    db.add(a)
+    await db.flush()
+    img = ImageRecord(
+        artist_id=a.id, path=str(tmp_path / "p.jpg"),
+        sha256="e" * 64, size_bytes=10, mime="image/jpeg",
+        origin="imported_filesystem",
+    )
+    db.add(img)
+    used = Tag(name="kept", kind=TagKind.general)
+    unused = Tag(name="prune", kind=TagKind.general)
+    db.add_all([used, unused])
+    await db.flush()
+    await db.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=used.id,
+    ))
+    await db.commit()
+
+    resp = await client.post(
+        "/api/admin/tags/prune-unused", json={"dry_run": True},
+    )
+    body = await resp.get_json()
+    assert body["count"] == 1
+    assert "prune" in body["sample_names"]
+
+
+@pytest.mark.asyncio
+async def test_prune_unused_commit_deletes_and_returns_count(client, db):
+    db.add(Tag(name="byebye", kind=TagKind.general))
+    await db.commit()
+
+    resp = await client.post(
+        "/api/admin/tags/prune-unused", json={"dry_run": False},
+    )
+    body = await resp.get_json()
+    assert body["deleted"] >= 1
+    assert "byebye" in body["sample_names"]

tests/test_api_migrate.py

@@ -31,17 +31,10 @@ async def client(app):
         yield c
 
 
-@pytest.mark.asyncio
+# Retired 2026-05-24 (FC-3h): `test_post_backup_returns_202_and_id`
-async def test_post_backup_returns_202_and_id(client, monkeypatch):
+# asserted the /api/migrate/backup endpoint, which was retired in FC-3h.
-    monkeypatch.setattr(
+# Backup is now a first-class feature at /api/system/backup/*;
-        "backend.app.api.migrate.run_migration",
+# coverage lives in tests/test_api_system_backup.py.
-        type("F", (), {"delay": lambda self, *a, **k: None})(),
-    )
-    resp = await client.post("/api/migrate/backup", json={})
-    assert resp.status_code == 202
-    body = await resp.get_json()
-    assert "run_id" in body
-    assert body["status"] == "pending"
 
 
 @pytest.mark.asyncio
@@ -62,10 +55,6 @@ async def test_post_ingest_rejects_missing_file(client):
 
 @pytest.mark.asyncio
 async def test_post_ingest_accepts_multipart_file(client, monkeypatch):
-    monkeypatch.setattr(
-        "backend.app.api.migrate._has_recent_pre_migration_backup",
-        lambda: True,  # pretend backup exists
-    )
     monkeypatch.setattr(
         "backend.app.api.migrate.run_migration",
         type("F", (), {"delay": lambda self, *a, **k: None})(),
@@ -83,32 +72,18 @@ async def test_post_ingest_accepts_multipart_file(client, monkeypatch):
     assert resp.status_code == 202
 
 
-@pytest.mark.asyncio
+# Retired 2026-05-24: `test_post_apply_without_backup_rejected` asserted
-async def test_post_apply_without_backup_rejected(client, monkeypatch):
+# the migration backup-gate rejected applies without a recent backup.
-    monkeypatch.setattr(
+# That gate was removed in commit 5535677 (_APPLY_KINDS is now empty,
-        "backend.app.api.migrate._has_recent_pre_migration_backup",
+# FC-3h will own backup as a first-class feature) so the test was
-        lambda: False,
+# testing dead behavior. Removed rather than rewritten.
-    )
-    export = {
-        "source_app": "gallerysubscriber", "schema_version": 1,
-        "subscriptions": [], "credentials": [],
-    }
-    resp = await client.post(
-        "/api/migrate/gs_ingest",
-        form={"dry_run": "false"},
-        files={"export_file": _file_storage(export, "gs-export.json")},
-    )
-    assert resp.status_code == 400
-    body = await resp.get_json()
-    assert body["error"] == "no_backup"
 
 
 @pytest.mark.asyncio
 async def test_post_dry_run_allowed_without_backup(client, monkeypatch):
-    monkeypatch.setattr(
+    # FC-3h: the _has_recent_pre_migration_backup gate was retired; dry-run
-        "backend.app.api.migrate._has_recent_pre_migration_backup",
+    # ingests were always allowed and now non-dry-run ingests are too.
-        lambda: False,
+    # This test stays as regression coverage that dry_run ingest still works.
-    )
     monkeypatch.setattr(
         "backend.app.api.migrate.run_migration",
         type("F", (), {"delay": lambda self, *a, **k: None})(),

tests/test_api_system_activity.py

@@ -0,0 +1,267 @@
+"""FC-3i: /api/system/activity/* endpoint tests.
+
+Mocks Redis LLEN + celery inspect via monkeypatch on the module-level
+_read_queues_sync / _read_workers_sync helpers, so tests don't need
+a live broker or running workers.
+"""
+
+from datetime import UTC, datetime, timedelta
+
+import pytest
+import pytest_asyncio
+
+from backend.app import create_app
+from backend.app.api import system_activity as activity_module
+from backend.app.models import TaskRun
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture
+async def app():
+    return create_app()
+
+
+@pytest.fixture
+async def client(app):
+    async with app.test_client() as c:
+        yield c
+
+
+@pytest.fixture(autouse=True)
+def _reset_caches(monkeypatch):
+    """Clear the module-level caches between tests so cache state from
+    one test doesn't leak into the next."""
+    monkeypatch.setitem(activity_module._QUEUE_CACHE, "data", None)
+    monkeypatch.setitem(activity_module._QUEUE_CACHE, "ts", 0.0)
+    monkeypatch.setitem(activity_module._WORKER_CACHE, "data", None)
+    monkeypatch.setitem(activity_module._WORKER_CACHE, "ts", 0.0)
+
+
+# --- /queues -------------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_queues_returns_all_known_queues(client, monkeypatch):
+    def _fake():
+        return {
+            "queues": dict.fromkeys(activity_module._QUEUE_NAMES, 0),
+            "fetched_at": datetime.now(UTC).isoformat(),
+        }
+    monkeypatch.setattr(activity_module, "_read_queues_sync", _fake)
+
+    resp = await client.get("/api/system/activity/queues")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert set(body["queues"].keys()) == set(activity_module._QUEUE_NAMES)
+
+
+@pytest.mark.asyncio
+async def test_queues_cached_within_ttl(client, monkeypatch):
+    call_count = {"n": 0}
+
+    def _counting():
+        call_count["n"] += 1
+        return {"queues": {}, "fetched_at": "x"}
+    monkeypatch.setattr(activity_module, "_read_queues_sync", _counting)
+
+    await client.get("/api/system/activity/queues")
+    await client.get("/api/system/activity/queues")
+    await client.get("/api/system/activity/queues")
+    assert call_count["n"] == 1  # cached after the first call
+
+
+@pytest.mark.asyncio
+async def test_queues_redis_failure_returns_null_per_failing_queue(client, monkeypatch):
+    """_read_queues_sync's per-queue try/except returns None on failure."""
+    def _partial():
+        out = dict.fromkeys(activity_module._QUEUE_NAMES, 0)
+        out["ml"] = None  # simulate broker hiccup on the ml queue
+        return {"queues": out, "fetched_at": "x"}
+    monkeypatch.setattr(activity_module, "_read_queues_sync", _partial)
+
+    resp = await client.get("/api/system/activity/queues")
+    body = await resp.get_json()
+    assert body["queues"]["ml"] is None
+    assert body["queues"]["import"] == 0
+
+
+# --- /workers ------------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_workers_returns_per_worker_queue_membership(client, monkeypatch):
+    def _fake():
+        return {
+            "workers": {
+                "worker@host1": {"queues": ["import", "thumbnail"], "active_count": 2},
+                "worker@host2": {"queues": ["ml"], "active_count": 0},
+            },
+            "fetched_at": "x",
+        }
+    monkeypatch.setattr(activity_module, "_read_workers_sync", _fake)
+
+    resp = await client.get("/api/system/activity/workers")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert "worker@host1" in body["workers"]
+    assert "import" in body["workers"]["worker@host1"]["queues"]
+
+
+@pytest.mark.asyncio
+async def test_workers_cached_within_ttl(client, monkeypatch):
+    call_count = {"n": 0}
+
+    def _counting():
+        call_count["n"] += 1
+        return {"workers": {}, "fetched_at": "x"}
+    monkeypatch.setattr(activity_module, "_read_workers_sync", _counting)
+
+    await client.get("/api/system/activity/workers")
+    await client.get("/api/system/activity/workers")
+    assert call_count["n"] == 1
+
+
+# --- /runs ---------------------------------------------------------
+
+
+@pytest_asyncio.fixture
+async def _seed_runs(db):
+    """Insert 5 task_run rows for paging tests."""
+    now = datetime.now(UTC)
+    for i in range(5):
+        db.add(TaskRun(
+            celery_task_id=f"tid-{i}",
+            queue="import" if i % 2 == 0 else "ml",
+            task_name=f"backend.app.tasks.fake.task_{i}",
+            target_id=i,
+            started_at=now - timedelta(seconds=10 - i),
+            finished_at=now - timedelta(seconds=9 - i),
+            duration_ms=1000,
+            status="ok" if i < 3 else "error",
+            error_type="ValueError" if i >= 3 else None,
+            error_message="boom" if i >= 3 else None,
+        ))
+    await db.commit()
+
+
+@pytest.mark.asyncio
+async def test_runs_paginated_descending_by_id(client, _seed_runs):
+    resp = await client.get("/api/system/activity/runs?limit=3")
+    body = await resp.get_json()
+    assert len(body["runs"]) == 3
+    # Descending: latest id first.
+    ids = [r["id"] for r in body["runs"]]
+    assert ids == sorted(ids, reverse=True)
+    assert body["next_cursor"] is not None  # 5 total, asked for 3 → more
+
+
+@pytest.mark.asyncio
+async def test_runs_filter_by_queue(client, _seed_runs):
+    resp = await client.get("/api/system/activity/runs?queue=ml")
+    body = await resp.get_json()
+    assert all(r["queue"] == "ml" for r in body["runs"])
+    assert len(body["runs"]) == 2  # i=1, i=3
+
+
+@pytest.mark.asyncio
+async def test_runs_filter_by_status(client, _seed_runs):
+    resp = await client.get("/api/system/activity/runs?status=error")
+    body = await resp.get_json()
+    assert all(r["status"] == "error" for r in body["runs"])
+    assert len(body["runs"]) == 2  # i=3, i=4
+
+
+@pytest.mark.asyncio
+async def test_runs_keyset_cursor(client, _seed_runs):
+    page1 = await (await client.get("/api/system/activity/runs?limit=2")).get_json()
+    assert len(page1["runs"]) == 2
+    assert page1["next_cursor"] is not None
+
+    page2 = await (await client.get(
+        f"/api/system/activity/runs?limit=2&before_id={page1['next_cursor']}"
+    )).get_json()
+    assert len(page2["runs"]) == 2
+    page1_ids = {r["id"] for r in page1["runs"]}
+    page2_ids = {r["id"] for r in page2["runs"]}
+    assert page1_ids.isdisjoint(page2_ids)
+
+
+@pytest.mark.asyncio
+async def test_runs_invalid_limit_400(client):
+    resp = await client.get("/api/system/activity/runs?limit=not-a-number")
+    assert resp.status_code == 400
+
+
+# --- /failures -----------------------------------------------------
+
+
+@pytest_asyncio.fixture
+async def _seed_failures(db):
+    """Mix of ok / error / timeout rows; some old, some recent."""
+    now = datetime.now(UTC)
+    recent = [
+        ("error", "OperationalError"),
+        ("error", "OperationalError"),
+        ("error", "OperationalError"),
+        ("timeout", "TimeoutError"),
+        ("timeout", "TimeoutError"),
+        ("error", "OSError"),
+        ("ok", None),  # success — should NOT appear in failures
+    ]
+    for i, (status, etype) in enumerate(recent):
+        db.add(TaskRun(
+            celery_task_id=f"f-{i}",
+            queue="ml",
+            task_name="backend.app.tasks.fake.x",
+            target_id=i,
+            started_at=now - timedelta(seconds=120 - i),
+            finished_at=now - timedelta(seconds=60 - i),
+            duration_ms=1000,
+            status=status,
+            error_type=etype,
+            error_message="boom" if status != "ok" else None,
+        ))
+    # One ancient failure (>24h) that should NOT appear.
+    db.add(TaskRun(
+        celery_task_id="f-old",
+        queue="ml",
+        task_name="backend.app.tasks.fake.x",
+        target_id=999,
+        started_at=now - timedelta(hours=30),
+        finished_at=now - timedelta(hours=29),
+        duration_ms=1000,
+        status="error",
+        error_type="OldError",
+        error_message="ancient",
+    ))
+    await db.commit()
+
+
+@pytest.mark.asyncio
+async def test_failures_returns_recent_errors_and_timeouts(client, _seed_failures):
+    resp = await client.get("/api/system/activity/failures")
+    body = await resp.get_json()
+    statuses = {r["status"] for r in body["recent"]}
+    assert statuses == {"error", "timeout"}
+    # No ok rows.
+    assert "ok" not in statuses
+
+
+@pytest.mark.asyncio
+async def test_failures_count_by_type_groups_correctly(client, _seed_failures):
+    resp = await client.get("/api/system/activity/failures")
+    body = await resp.get_json()
+    counts = body["count_by_type"]
+    assert counts.get("OperationalError") == 3
+    assert counts.get("TimeoutError") == 2
+    assert counts.get("OSError") == 1
+    assert "OldError" not in counts  # outside 24h window
+
+
+@pytest.mark.asyncio
+async def test_failures_only_within_24h_window(client, _seed_failures):
+    resp = await client.get("/api/system/activity/failures")
+    body = await resp.get_json()
+    ids = {r["error_type"] for r in body["recent"]}
+    assert "OldError" not in ids

tests/test_api_system_backup.py

@@ -0,0 +1,336 @@
+"""FC-3h: /api/system/backup/* endpoint integration tests."""
+from datetime import UTC, datetime, timedelta
+
+import pytest
+import pytest_asyncio
+from sqlalchemy import select
+
+from backend.app import create_app
+from backend.app.models import BackupRun
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture
+async def app():
+    return create_app()
+
+
+@pytest.fixture
+async def client(app):
+    async with app.test_client() as c:
+        yield c
+
+
+@pytest_asyncio.fixture
+async def _seed_runs(db):
+    """Insert 4 BackupRun rows for paging/filter tests."""
+    now = datetime.now(UTC)
+    for i in range(4):
+        db.add(BackupRun(
+            kind="db" if i % 2 == 0 else "images",
+            status="ok",
+            tag=None,
+            triggered_by="manual",
+            started_at=now - timedelta(seconds=10 - i),
+            finished_at=now - timedelta(seconds=9 - i),
+            sql_path="/tmp/fake.sql" if i % 2 == 0 else None,
+            tar_path=None if i % 2 == 0 else "/tmp/fake.tar.zst",
+            size_bytes=100,
+            manifest={},
+        ))
+    await db.commit()
+
+
+# --- POST /db, /images ----------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_post_db_dispatches_and_returns_202(client, monkeypatch):
+    dispatched = []
+    monkeypatch.setattr(
+        "backend.app.tasks.backup.backup_db_task.delay",
+        lambda **kw: dispatched.append(kw),
+    )
+    resp = await client.post("/api/system/backup/db", json={})
+    assert resp.status_code == 202
+    body = await resp.get_json()
+    assert body["status"] == "dispatched"
+    assert dispatched == [{"tag": None, "triggered_by": "manual"}]
+
+
+@pytest.mark.asyncio
+async def test_post_db_persists_tag(client, monkeypatch):
+    dispatched = []
+    monkeypatch.setattr(
+        "backend.app.tasks.backup.backup_db_task.delay",
+        lambda **kw: dispatched.append(kw),
+    )
+    resp = await client.post(
+        "/api/system/backup/db", json={"tag": "pre-cutover"},
+    )
+    assert resp.status_code == 202
+    assert dispatched[0]["tag"] == "pre-cutover"
+
+
+@pytest.mark.asyncio
+async def test_post_db_rejects_tag_too_long(client):
+    resp = await client.post(
+        "/api/system/backup/db", json={"tag": "x" * 65},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "invalid_tag"
+
+
+@pytest.mark.asyncio
+async def test_post_images_dispatches(client, monkeypatch):
+    dispatched = []
+    monkeypatch.setattr(
+        "backend.app.tasks.backup.backup_images_task.delay",
+        lambda **kw: dispatched.append(kw),
+    )
+    resp = await client.post("/api/system/backup/images", json={})
+    assert resp.status_code == 202
+
+
+# --- GET /runs -------------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_runs_paginated_descending(client, _seed_runs):
+    resp = await client.get("/api/system/backup/runs?limit=2")
+    body = await resp.get_json()
+    assert len(body["runs"]) == 2
+    ids = [r["id"] for r in body["runs"]]
+    assert ids == sorted(ids, reverse=True)
+    assert body["next_cursor"] is not None
+
+
+@pytest.mark.asyncio
+async def test_runs_filter_by_kind_db(client, _seed_runs):
+    resp = await client.get("/api/system/backup/runs?kind=db")
+    body = await resp.get_json()
+    assert all(r["kind"] == "db" for r in body["runs"])
+
+
+@pytest.mark.asyncio
+async def test_runs_filter_by_kind_images(client, _seed_runs):
+    resp = await client.get("/api/system/backup/runs?kind=images")
+    body = await resp.get_json()
+    assert all(r["kind"] == "images" for r in body["runs"])
+
+
+@pytest.mark.asyncio
+async def test_runs_rejects_invalid_kind(client):
+    resp = await client.get("/api/system/backup/runs?kind=bogus")
+    assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+async def test_runs_rejects_invalid_limit(client):
+    resp = await client.get("/api/system/backup/runs?limit=not-int")
+    assert resp.status_code == 400
+
+
+# --- GET /runs/<id> --------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_get_run_returns_row(client, _seed_runs):
+    list_resp = await client.get("/api/system/backup/runs?limit=1")
+    rid = (await list_resp.get_json())["runs"][0]["id"]
+    resp = await client.get(f"/api/system/backup/runs/{rid}")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["id"] == rid
+
+
+@pytest.mark.asyncio
+async def test_get_run_404(client):
+    resp = await client.get("/api/system/backup/runs/999999")
+    assert resp.status_code == 404
+
+
+# --- PATCH /runs/<id> ------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_patch_tag_sets_value(client, _seed_runs, db_sync):
+    list_resp = await client.get("/api/system/backup/runs?limit=1")
+    rid = (await list_resp.get_json())["runs"][0]["id"]
+
+    resp = await client.patch(
+        f"/api/system/backup/runs/{rid}",
+        json={"tag": "monthly"},
+    )
+    assert resp.status_code == 200
+
+    tag = db_sync.execute(
+        select(BackupRun.tag).where(BackupRun.id == rid)
+    ).scalar_one()
+    assert tag == "monthly"
+
+
+@pytest.mark.asyncio
+async def test_patch_tag_null_clears(client, _seed_runs, db_sync):
+    list_resp = await client.get("/api/system/backup/runs?limit=1")
+    rid = (await list_resp.get_json())["runs"][0]["id"]
+
+    await client.patch(f"/api/system/backup/runs/{rid}", json={"tag": "x"})
+    await client.patch(f"/api/system/backup/runs/{rid}", json={"tag": None})
+
+    tag = db_sync.execute(
+        select(BackupRun.tag).where(BackupRun.id == rid)
+    ).scalar_one()
+    assert tag is None
+
+
+# --- POST /runs/<id>/restore ----------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_restore_wrong_confirm_400(client, _seed_runs):
+    list_resp = await client.get("/api/system/backup/runs?kind=db&limit=1")
+    rid = (await list_resp.get_json())["runs"][0]["id"]
+
+    resp = await client.post(
+        f"/api/system/backup/runs/{rid}/restore",
+        json={"confirm": "wrong"},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "confirm_mismatch"
+    assert body["expected"] == f"restore-db-{rid}"
+
+
+@pytest.mark.asyncio
+async def test_restore_correct_confirm_dispatches(client, _seed_runs, monkeypatch):
+    list_resp = await client.get("/api/system/backup/runs?kind=db&limit=1")
+    rid = (await list_resp.get_json())["runs"][0]["id"]
+
+    dispatched = []
+    monkeypatch.setattr(
+        "backend.app.tasks.backup.restore_db_task.delay",
+        lambda **kw: dispatched.append(kw),
+    )
+
+    resp = await client.post(
+        f"/api/system/backup/runs/{rid}/restore",
+        json={"confirm": f"restore-db-{rid}"},
+    )
+    assert resp.status_code == 202
+    assert dispatched == [{"source_backup_run_id": rid}]
+
+
+@pytest.mark.asyncio
+async def test_restore_rejects_non_ok_status(client, db):
+    # Seed an error-status row directly.
+    db.add(BackupRun(
+        kind="db", status="error", tag=None, triggered_by="manual",
+        started_at=datetime.now(UTC), finished_at=datetime.now(UTC),
+        sql_path="/tmp/fake.sql", manifest={},
+    ))
+    await db.commit()
+    rid_q = await client.get("/api/system/backup/runs?kind=db&limit=1")
+    rid = (await rid_q.get_json())["runs"][0]["id"]
+
+    resp = await client.post(
+        f"/api/system/backup/runs/{rid}/restore",
+        json={"confirm": f"restore-db-{rid}"},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "not_restorable"
+
+
+# --- DELETE /runs/<id> ----------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_delete_wrong_confirm_400(client, _seed_runs):
+    list_resp = await client.get("/api/system/backup/runs?limit=1")
+    rid = (await list_resp.get_json())["runs"][0]["id"]
+
+    resp = await client.delete(
+        f"/api/system/backup/runs/{rid}",
+        json={"confirm": "wrong"},
+    )
+    assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+async def test_delete_correct_confirm_204(client, _seed_runs, db_sync):
+    list_resp = await client.get("/api/system/backup/runs?limit=1")
+    row = (await list_resp.get_json())["runs"][0]
+    rid = row["id"]
+    kind = row["kind"]
+
+    resp = await client.delete(
+        f"/api/system/backup/runs/{rid}",
+        json={"confirm": f"delete-{kind}-{rid}"},
+    )
+    assert resp.status_code == 204
+
+    surviving = db_sync.execute(
+        select(BackupRun.id).where(BackupRun.id == rid)
+    ).scalar_one_or_none()
+    assert surviving is None
+
+
+# --- /settings ------------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_get_settings_returns_defaults(client):
+    resp = await client.get("/api/system/backup/settings")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["backup_db_nightly_enabled"] is False
+    assert body["backup_db_nightly_hour_utc"] == 3
+    assert body["backup_db_keep_last_n"] == 14
+    assert body["backup_images_keep_last_n"] == 3
+
+
+@pytest.mark.asyncio
+async def test_patch_settings_updates_values(client):
+    resp = await client.patch(
+        "/api/system/backup/settings",
+        json={
+            "backup_db_nightly_enabled": True,
+            "backup_db_nightly_hour_utc": 5,
+            "backup_db_keep_last_n": 30,
+        },
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["backup_db_nightly_enabled"] is True
+    assert body["backup_db_nightly_hour_utc"] == 5
+    assert body["backup_db_keep_last_n"] == 30
+
+
+@pytest.mark.asyncio
+async def test_patch_settings_rejects_invalid_bool(client):
+    resp = await client.patch(
+        "/api/system/backup/settings",
+        json={"backup_db_nightly_enabled": "yes"},
+    )
+    assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+async def test_patch_settings_rejects_hour_out_of_range(client):
+    resp = await client.patch(
+        "/api/system/backup/settings",
+        json={"backup_db_nightly_hour_utc": 24},
+    )
+    assert resp.status_code == 400
+
+
+@pytest.mark.asyncio
+async def test_patch_settings_rejects_keep_n_below_min(client):
+    resp = await client.patch(
+        "/api/system/backup/settings",
+        json={"backup_db_keep_last_n": 0},
+    )
+    assert resp.status_code == 400

tests/test_backup_service.py

@@ -0,0 +1,200 @@
+"""FC-3h: backup_service unit tests.
+
+Subprocess calls (pg_dump, tar, psql) are monkeypatched so tests run
+without external binaries. The real subprocess behavior is exercised
+implicitly via the Celery task tests in test_tasks_backup.py.
+"""
+import json
+from pathlib import Path
+
+import pytest
+
+from backend.app.services import backup_service
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture
+def fake_subprocess(monkeypatch):
+    """Replace subprocess.run with a fake that writes a sentinel to
+    the target path (for pg_dump's -f, for tar's -cf). Captures all
+    calls in a list."""
+    calls = []
+
+    class _FakeProc:
+        returncode = 0
+        stdout = b""
+        stderr = b""
+
+    def _fake_run(cmd, **kwargs):
+        calls.append(list(cmd))
+        if cmd[0] == "pg_dump":
+            i = cmd.index("-f")
+            Path(cmd[i + 1]).write_bytes(b"-- fake pg_dump\n")
+        elif cmd[0] == "tar" and "-cf" in cmd:
+            i = cmd.index("-cf")
+            Path(cmd[i + 1]).write_bytes(b"fake tar payload")
+        return _FakeProc()
+
+    monkeypatch.setattr("subprocess.run", _fake_run)
+    return calls
+
+
+# --- backup_db -------------------------------------------------------
+
+
+def test_backup_db_writes_sql_and_manifest(tmp_path, fake_subprocess):
+    result = backup_service.backup_db(
+        db_url="postgresql://test@x/db",
+        images_root=tmp_path,
+        tag="pre-cutover",
+        triggered_by="manual",
+    )
+    sql = Path(result["sql_path"])
+    manifest = Path(result["manifest_path"])
+    assert sql.is_file() and sql.suffix == ".sql"
+    assert manifest.is_file() and manifest.suffix == ".json"
+    assert result["kind"] == "db"
+    assert result["tar_path"] is None
+    assert result["size_bytes"] == len(b"-- fake pg_dump\n")
+
+    parsed = json.loads(manifest.read_text())
+    assert parsed["kind"] == "db"
+    assert parsed["tag"] == "pre-cutover"
+    assert parsed["triggered_by"] == "manual"
+    assert parsed["artifact_path"] == str(sql)
+
+
+def test_backup_db_strips_sqlalchemy_psycopg_driver(tmp_path, fake_subprocess):
+    backup_service.backup_db(
+        db_url="postgresql+psycopg://u:p@h/d", images_root=tmp_path,
+    )
+    cmd = fake_subprocess[0]
+    assert cmd[0] == "pg_dump"
+    assert cmd[-1].startswith("postgresql://")
+    assert "+psycopg" not in cmd[-1]
+
+
+def test_backup_db_strips_asyncpg_driver(tmp_path, fake_subprocess):
+    backup_service.backup_db(
+        db_url="postgresql+asyncpg://u:p@h/d", images_root=tmp_path,
+    )
+    assert "+asyncpg" not in fake_subprocess[0][-1]
+
+
+def test_backup_db_default_tag_is_none(tmp_path, fake_subprocess):
+    result = backup_service.backup_db(
+        db_url="postgresql://u@h/d", images_root=tmp_path,
+    )
+    parsed = json.loads(Path(result["manifest_path"]).read_text())
+    assert parsed["tag"] is None
+
+
+# --- backup_images ---------------------------------------------------
+
+
+def test_backup_images_writes_tar_and_manifest(tmp_path, fake_subprocess):
+    result = backup_service.backup_images(
+        images_root=tmp_path, tag="monthly", triggered_by="manual",
+    )
+    tar = Path(result["tar_path"])
+    assert tar.is_file() and tar.name.endswith(".tar.zst")
+    assert result["sql_path"] is None
+    assert result["size_bytes"] == len(b"fake tar payload")
+
+
+def test_backup_images_excludes_backups_and_quarantine(tmp_path, fake_subprocess):
+    backup_service.backup_images(images_root=tmp_path)
+    cmd = fake_subprocess[0]
+    excludes = [arg for arg in cmd if arg.startswith("--exclude=")]
+    assert any("_backups" in e for e in excludes)
+    assert any("_quarantine" in e for e in excludes)
+
+
+# --- restore_db ------------------------------------------------------
+
+
+def test_restore_db_drops_schema_then_loads(tmp_path, fake_subprocess):
+    sql_path = tmp_path / "fake.sql"
+    sql_path.write_text("SELECT 1;")
+    backup_service.restore_db(
+        db_url="postgresql://u@h/d", sql_path=sql_path,
+    )
+    # Two psql calls: one with -c (DROP SCHEMA), one with -f (load).
+    assert len(fake_subprocess) == 2
+    assert "-c" in fake_subprocess[0]
+    assert "DROP SCHEMA IF EXISTS public CASCADE" in fake_subprocess[0][-1]
+    assert "-f" in fake_subprocess[1]
+    assert str(sql_path) in fake_subprocess[1]
+
+
+# --- restore_images --------------------------------------------------
+
+
+def test_restore_images_untar_to_parent(tmp_path, fake_subprocess):
+    tar_path = tmp_path / "fake.tar.zst"
+    tar_path.write_bytes(b"fake")
+    backup_service.restore_images(images_root=tmp_path, tar_path=tar_path)
+    cmd = fake_subprocess[0]
+    assert cmd[:3] == ["tar", "--zstd", "-xf"]
+    assert str(tar_path) in cmd
+    assert "-C" in cmd
+
+
+# --- unlink ----------------------------------------------------------
+
+
+def test_unlink_removes_present_files_and_reports(tmp_path):
+    sql = tmp_path / "x.sql"
+    sql.write_bytes(b"x")
+    tar = tmp_path / "x.tar.zst"
+    tar.write_bytes(b"x")
+    manifest = tmp_path / "x.json"
+    manifest.write_text("{}")
+    result = backup_service.unlink_artifact_files(
+        sql_path=str(sql), tar_path=str(tar), manifest_path=str(manifest),
+    )
+    assert result == {"sql": True, "tar": True, "manifest": True}
+    assert not sql.exists() and not tar.exists() and not manifest.exists()
+
+
+def test_unlink_missing_files_returns_true(tmp_path):
+    """missing_ok semantics: a non-existent file isn't an error."""
+    result = backup_service.unlink_artifact_files(
+        sql_path=str(tmp_path / "nope.sql"),
+        tar_path=None, manifest_path=None,
+    )
+    assert result == {"sql": True}
+
+
+def test_unlink_skips_none_paths():
+    """A None path is skipped — not added to the result dict."""
+    result = backup_service.unlink_artifact_files(
+        sql_path=None, tar_path=None, manifest_path=None,
+    )
+    assert result == {}
+
+
+# --- helpers ---------------------------------------------------------
+
+
+def test_libpq_url_strips_each_known_driver():
+    assert backup_service._libpq_url(
+        "postgresql+psycopg://u@h/d"
+    ) == "postgresql://u@h/d"
+    assert backup_service._libpq_url(
+        "postgresql+asyncpg://u@h/d"
+    ) == "postgresql://u@h/d"
+    assert backup_service._libpq_url(
+        "postgresql+psycopg2://u@h/d"
+    ) == "postgresql://u@h/d"
+    # Plain URL passes through unchanged.
+    assert backup_service._libpq_url(
+        "postgresql://u@h/d"
+    ) == "postgresql://u@h/d"
+
+
+def test_backups_dir_created_on_first_use(tmp_path):
+    d = backup_service._backups_dir(tmp_path)
+    assert d.is_dir()
+    assert d.name == "_backups"

tests/test_cleanup_service.py

@@ -0,0 +1,335 @@
+"""FC-3k: cleanup_service unit tests.
+
+Mutations go against real Postgres (db_sync fixture). File-system
+side effects use tmp_path. Assertions on mutated rows use COLUMN
+SELECTS per reference_async_coredml_test_assertions — never
+re-read ORM attributes after a service mutates and re-fetches.
+"""
+import pytest
+from sqlalchemy import func, select
+
+from backend.app.models import Artist, ImageRecord, Tag, TagKind
+from backend.app.models.tag import image_tag
+from backend.app.services import cleanup_service
+
+pytestmark = pytest.mark.integration
+
+
+def _make_image(db_sync, *, artist, path, sha256, size=1000, thumb=None):
+    img = ImageRecord(
+        artist_id=artist.id,
+        path=path,
+        sha256=sha256,
+        size_bytes=size,
+        mime="image/jpeg",
+        origin="imported_filesystem",
+        thumbnail_path=thumb,
+    )
+    db_sync.add(img)
+    db_sync.flush()
+    return img
+
+
+def _make_artist(db_sync, *, slug="aria", name="Aria"):
+    a = Artist(name=name, slug=slug)
+    db_sync.add(a)
+    db_sync.flush()
+    return a
+
+
+def _make_tag(db_sync, *, name, kind=TagKind.general):
+    t = Tag(name=name, kind=kind)
+    db_sync.add(t)
+    db_sync.flush()
+    return t
+
+
+# --- project_artist_cascade -----------------------------------------
+
+
+def test_project_artist_cascade_returns_zeroes_for_empty_artist(db_sync):
+    _make_artist(db_sync, slug="empty")
+    db_sync.commit()
+    result = cleanup_service.project_artist_cascade(db_sync, slug="empty")
+    assert result["artist"]["slug"] == "empty"
+    assert result["projected"] == {
+        "images": 0,
+        "sources": 0,
+        "thumbs": 0,
+        "import_tasks": 0,
+        "bytes_on_disk": 0,
+    }
+
+
+def test_project_artist_cascade_counts_images_and_thumbs_and_bytes(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="counted")
+    _make_image(
+        db_sync, artist=a, path=str(tmp_path / "a.jpg"),
+        sha256="a" * 64, size=1000, thumb=str(tmp_path / "a.thumb"),
+    )
+    _make_image(
+        db_sync, artist=a, path=str(tmp_path / "b.jpg"),
+        sha256="b" * 64, size=2500, thumb=None,
+    )
+    db_sync.commit()
+    result = cleanup_service.project_artist_cascade(db_sync, slug="counted")
+    assert result["projected"]["images"] == 2
+    assert result["projected"]["thumbs"] == 1
+    assert result["projected"]["bytes_on_disk"] == 3500
+
+
+def test_project_artist_cascade_raises_on_unknown_slug(db_sync):
+    with pytest.raises(LookupError):
+        cleanup_service.project_artist_cascade(db_sync, slug="nope")
+
+
+# --- project_bulk_image_delete --------------------------------------
+
+
+def test_project_bulk_image_delete_separates_found_from_missing(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="bd")
+    i1 = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "1.jpg"),
+        sha256="1" * 64, size=10, thumb="t1",
+    )
+    i2 = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "2.jpg"),
+        sha256="2" * 64, size=20, thumb=None,
+    )
+    db_sync.commit()
+    result = cleanup_service.project_bulk_image_delete(
+        db_sync, image_ids=[i1.id, i2.id, 9_999_999],
+    )
+    assert result["images_found"] == 2
+    assert result["thumbs_to_unlink"] == 1
+    assert result["bytes_on_disk"] == 30
+    assert result["missing_ids"] == [9_999_999]
+
+
+def test_project_bulk_image_delete_empty_input(db_sync):
+    result = cleanup_service.project_bulk_image_delete(db_sync, image_ids=[])
+    assert result == {
+        "images_found": 0,
+        "thumbs_to_unlink": 0,
+        "bytes_on_disk": 0,
+        "missing_ids": [],
+    }
+
+
+# --- count_tag_associations / find_unused_tags ----------------------
+
+
+def test_count_tag_associations_counts_image_tag_rows(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="ta")
+    img = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "x.jpg"), sha256="c" * 64,
+    )
+    tag = _make_tag(db_sync, name="cyberpunk")
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=tag.id,
+    ))
+    db_sync.commit()
+    assert cleanup_service.count_tag_associations(db_sync, tag_id=tag.id) == 1
+
+
+def test_find_unused_tags_returns_only_unreferenced(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="fu")
+    img = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "y.jpg"), sha256="d" * 64,
+    )
+    used = _make_tag(db_sync, name="used")
+    _make_tag(db_sync, name="aaa-unused")
+    _make_tag(db_sync, name="zzz-unused")
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=used.id,
+    ))
+    db_sync.commit()
+    result = cleanup_service.find_unused_tags(db_sync)
+    names = [t.name for t in result]
+    assert "used" not in names
+    assert "aaa-unused" in names
+    assert "zzz-unused" in names
+    # Sorted by name ascending.
+    assert names.index("aaa-unused") < names.index("zzz-unused")
+
+
+# --- unlink_image_files ---------------------------------------------
+
+
+def test_unlink_image_files_removes_original_and_thumbnail(db_sync, tmp_path):
+    original = tmp_path / "orig.jpg"
+    original.write_bytes(b"x")
+    custom_thumb = tmp_path / "custom.thumb"
+    custom_thumb.write_bytes(b"x")
+    conv_thumb = tmp_path / "thumbs" / "aaa" / ("a" * 64 + ".jpg")
+    conv_thumb.parent.mkdir(parents=True)
+    conv_thumb.write_bytes(b"x")
+
+    img = ImageRecord(
+        artist_id=None, path=str(original), sha256="a" * 64,
+        size_bytes=1, thumbnail_path=str(custom_thumb),
+    )
+    result = cleanup_service.unlink_image_files(img, tmp_path)
+    assert result == {"original": True, "thumbnail": True}
+    assert not original.exists()
+    assert not custom_thumb.exists()
+    assert not conv_thumb.exists()
+
+
+def test_unlink_image_files_missing_files_count_as_success(tmp_path):
+    img = ImageRecord(
+        artist_id=None, path=str(tmp_path / "nope.jpg"),
+        sha256="b" * 64, size_bytes=1, thumbnail_path=None,
+    )
+    result = cleanup_service.unlink_image_files(img, tmp_path)
+    assert result == {"original": True, "thumbnail": False}
+
+
+# --- delete_artist_cascade ------------------------------------------
+
+
+def test_delete_artist_cascade_removes_images_and_artist_row(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="cas")
+    for i in range(3):
+        f = tmp_path / f"img{i}.jpg"
+        f.write_bytes(b"x")
+        _make_image(
+            db_sync, artist=a, path=str(f),
+            sha256=f"{i:064x}", size=10,
+        )
+    db_sync.commit()
+    artist_id = a.id
+
+    result = cleanup_service.delete_artist_cascade(
+        db_sync, artist_id=artist_id, images_root=tmp_path,
+    )
+    assert result["artist"]["slug"] == "cas"
+    assert result["summary"]["images_deleted"] == 3
+    assert result["summary"]["files_deleted"] == 3
+
+    # Column-select assertions per reference_async_coredml_test_assertions.
+    surviving_artist = db_sync.execute(
+        select(func.count(Artist.id)).where(Artist.id == artist_id)
+    ).scalar_one()
+    surviving_images = db_sync.execute(
+        select(func.count(ImageRecord.id))
+        .where(ImageRecord.artist_id == artist_id)
+    ).scalar_one()
+    assert surviving_artist == 0
+    assert surviving_images == 0
+
+
+def test_delete_artist_cascade_idempotent_on_missing(db_sync, tmp_path):
+    result = cleanup_service.delete_artist_cascade(
+        db_sync, artist_id=9_999_999, images_root=tmp_path,
+    )
+    assert result["artist"] is None
+    assert result["summary"]["images_deleted"] == 0
+
+
+# --- delete_images --------------------------------------------------
+
+
+def test_delete_images_removes_rows_and_files(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="di")
+    f1 = tmp_path / "1.jpg"
+    f1.write_bytes(b"x")
+    f2 = tmp_path / "2.jpg"
+    f2.write_bytes(b"x")
+    i1 = _make_image(db_sync, artist=a, path=str(f1), sha256="1" * 64)
+    i2 = _make_image(db_sync, artist=a, path=str(f2), sha256="2" * 64)
+    db_sync.commit()
+
+    result = cleanup_service.delete_images(
+        db_sync, image_ids=[i1.id, i2.id, 9_999_999],
+        images_root=tmp_path,
+    )
+    assert result["images_deleted"] == 2
+    assert result["files_deleted"] == 2
+    assert result["missing_ids"] == [9_999_999]
+
+    surviving = db_sync.execute(
+        select(func.count(ImageRecord.id))
+        .where(ImageRecord.id.in_([i1.id, i2.id]))
+    ).scalar_one()
+    assert surviving == 0
+
+
+# --- delete_tag -----------------------------------------------------
+
+
+def test_delete_tag_cascades_associations(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="dt")
+    img = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "x.jpg"), sha256="d" * 64,
+    )
+    tag = _make_tag(db_sync, name="doomed")
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=tag.id,
+    ))
+    db_sync.commit()
+    tag_id = tag.id
+
+    result = cleanup_service.delete_tag(db_sync, tag_id=tag_id)
+    assert result["deleted"]["id"] == tag_id
+    assert result["associations_removed"] == 1
+
+    surviving_tag = db_sync.execute(
+        select(func.count(Tag.id)).where(Tag.id == tag_id)
+    ).scalar_one()
+    surviving_assoc = db_sync.execute(
+        select(func.count())
+        .select_from(image_tag).where(image_tag.c.tag_id == tag_id)
+    ).scalar_one()
+    assert surviving_tag == 0
+    assert surviving_assoc == 0
+
+
+def test_delete_tag_raises_on_unknown_id(db_sync):
+    with pytest.raises(LookupError):
+        cleanup_service.delete_tag(db_sync, tag_id=9_999_999)
+
+
+# --- prune_unused_tags ----------------------------------------------
+
+
+def test_prune_unused_tags_dry_run_returns_count_and_names(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="pu")
+    img = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "z.jpg"), sha256="e" * 64,
+    )
+    used = _make_tag(db_sync, name="kept")
+    _make_tag(db_sync, name="prune-me-1")
+    _make_tag(db_sync, name="prune-me-2")
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=used.id,
+    ))
+    db_sync.commit()
+
+    result = cleanup_service.prune_unused_tags(db_sync, dry_run=True)
+    assert result["count"] == 2
+    assert "prune-me-1" in result["sample_names"]
+    assert "prune-me-2" in result["sample_names"]
+    # Nothing actually deleted.
+    surviving = db_sync.execute(select(func.count(Tag.id))).scalar_one()
+    assert surviving == 3
+
+
+def test_prune_unused_tags_commit_deletes_them(db_sync, tmp_path):
+    a = _make_artist(db_sync, slug="pu2")
+    img = _make_image(
+        db_sync, artist=a, path=str(tmp_path / "k.jpg"), sha256="f" * 64,
+    )
+    used = _make_tag(db_sync, name="kept")
+    _make_tag(db_sync, name="bye")
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=used.id,
+    ))
+    db_sync.commit()
+
+    result = cleanup_service.prune_unused_tags(db_sync, dry_run=False)
+    assert result["deleted"] == 1
+
+    surviving_names = db_sync.execute(select(Tag.name)).scalars().all()
+    assert "kept" in surviving_names
+    assert "bye" not in surviving_names

tests/test_maintenance.py

@@ -93,3 +93,174 @@ def test_cleanup_old_deletes_finished_old(db_sync):
 
     remaining = {t.source_path for t in db_sync.query(ImportTask).all()}
     assert remaining == {"/import/b.jpg", "/import/c.jpg"}
+
+
+# --- FC-3i: task_run sweep + retention -----------------------------
+
+
+def _make_task_run(db_sync, *, status, started_at, finished_at=None,
+                   error_type=None):
+    from backend.app.models import TaskRun
+    row = TaskRun(
+        celery_task_id="x",
+        queue="ml",
+        task_name="backend.app.tasks.fake.t",
+        target_id=1,
+        started_at=started_at,
+        finished_at=finished_at,
+        duration_ms=1000 if finished_at else None,
+        status=status,
+        error_type=error_type,
+        error_message="x" if status in ("error", "timeout") else None,
+    )
+    db_sync.add(row)
+    db_sync.flush()
+    return row.id
+
+
+def test_recover_stalled_task_runs_flips_old_running_to_error(db_sync):
+    from sqlalchemy import select
+
+    from backend.app.models import TaskRun
+    from backend.app.tasks.maintenance import recover_stalled_task_runs
+
+    now = datetime.now(UTC)
+    stale_id = _make_task_run(
+        db_sync, status="running", started_at=now - timedelta(minutes=10),
+    )
+    db_sync.commit()
+
+    recovered = recover_stalled_task_runs.apply().get()
+    assert recovered == 1
+
+    db_sync.expire_all()
+    status = db_sync.execute(
+        select(TaskRun.status).where(TaskRun.id == stale_id)
+    ).scalar_one()
+    error_type = db_sync.execute(
+        select(TaskRun.error_type).where(TaskRun.id == stale_id)
+    ).scalar_one()
+    assert status == "error"
+    assert error_type == "RecoverySweep"
+
+
+def test_recover_stalled_task_runs_skips_fresh_running(db_sync):
+    from sqlalchemy import select
+
+    from backend.app.models import TaskRun
+    from backend.app.tasks.maintenance import recover_stalled_task_runs
+
+    now = datetime.now(UTC)
+    fresh_id = _make_task_run(
+        db_sync, status="running", started_at=now - timedelta(seconds=30),
+    )
+    db_sync.commit()
+
+    recovered = recover_stalled_task_runs.apply().get()
+    assert recovered == 0
+
+    db_sync.expire_all()
+    status = db_sync.execute(
+        select(TaskRun.status).where(TaskRun.id == fresh_id)
+    ).scalar_one()
+    assert status == "running"
+
+
+def test_prune_task_runs_deletes_ok_older_than_24h(db_sync):
+    from sqlalchemy import select
+
+    from backend.app.models import TaskRun
+    from backend.app.tasks.maintenance import prune_task_runs
+
+    now = datetime.now(UTC)
+    old_id = _make_task_run(
+        db_sync, status="ok",
+        started_at=now - timedelta(hours=30),
+        finished_at=now - timedelta(hours=29),
+    )
+    recent_id = _make_task_run(
+        db_sync, status="ok",
+        started_at=now - timedelta(hours=2),
+        finished_at=now - timedelta(hours=1),
+    )
+    db_sync.commit()
+
+    result = prune_task_runs.apply().get()
+    assert result["ok_deleted"] == 1
+
+    db_sync.expire_all()
+    surviving_ids = set(db_sync.execute(
+        select(TaskRun.id).where(TaskRun.id.in_([old_id, recent_id]))
+    ).scalars().all())
+    assert surviving_ids == {recent_id}
+
+
+def test_prune_task_runs_deletes_failures_older_than_7d(db_sync):
+    from sqlalchemy import select
+
+    from backend.app.models import TaskRun
+    from backend.app.tasks.maintenance import prune_task_runs
+
+    now = datetime.now(UTC)
+    old_id = _make_task_run(
+        db_sync, status="error", error_type="OldError",
+        started_at=now - timedelta(days=10),
+        finished_at=now - timedelta(days=9),
+    )
+    db_sync.commit()
+
+    result = prune_task_runs.apply().get()
+    assert result["failures_deleted"] >= 1
+
+    db_sync.expire_all()
+    surviving = db_sync.execute(
+        select(TaskRun.id).where(TaskRun.id == old_id)
+    ).scalar_one_or_none()
+    assert surviving is None
+
+
+def test_prune_task_runs_keeps_recent_failures(db_sync):
+    from sqlalchemy import select
+
+    from backend.app.models import TaskRun
+    from backend.app.tasks.maintenance import prune_task_runs
+
+    now = datetime.now(UTC)
+    recent_id = _make_task_run(
+        db_sync, status="error", error_type="RecentError",
+        started_at=now - timedelta(days=6),
+        finished_at=now - timedelta(days=5),
+    )
+    db_sync.commit()
+
+    prune_task_runs.apply().get()
+
+    db_sync.expire_all()
+    surviving = db_sync.execute(
+        select(TaskRun.id).where(TaskRun.id == recent_id)
+    ).scalar_one_or_none()
+    assert surviving == recent_id
+
+
+def test_prune_task_runs_never_deletes_running(db_sync):
+    """Even a 30-day-old running row stays — recovery sweep is the
+    mechanism that flips them; prune doesn't touch in-flight."""
+    from sqlalchemy import select
+
+    from backend.app.models import TaskRun
+    from backend.app.tasks.maintenance import prune_task_runs
+
+    now = datetime.now(UTC)
+    ancient_id = _make_task_run(
+        db_sync, status="running",
+        started_at=now - timedelta(days=30),
+    )
+    db_sync.commit()
+
+    prune_task_runs.apply().get()
+
+    db_sync.expire_all()
+    surviving = db_sync.execute(
+        select(TaskRun.id).where(TaskRun.id == ancient_id)
+    ).scalar_one_or_none()
+    assert surviving == ancient_id

tests/test_migration_backup.py

@@ -1,64 +0,0 @@
-"""FC-5: backup + rollback service tests.
-
-Uses tmp_path to avoid touching real /images/_backups/. Subprocess
-calls (pg_dump, tar) are monkeypatched — the real shell-out is exercised
-in operator-side smoke testing on the homelab, not in unit tests.
-"""
-import pytest
-
-from backend.app.services.migrators import backup as backup_mod
-
-pytestmark = pytest.mark.integration
-
-
-@pytest.mark.asyncio
-async def test_create_backup_writes_sql_and_tar(monkeypatch, tmp_path):
-    calls = []
-
-    def fake_run(cmd, **kwargs):
-        calls.append(cmd)
-        # Touch the expected output file so existence checks pass downstream.
-        if "pg_dump" in cmd[0]:
-            outpath = tmp_path / "_backups" / f"fc_{kwargs['_test_ts']}.sql"
-            outpath.write_text("-- fake pg_dump output")
-        else:
-            outpath = tmp_path / "_backups" / f"fc_{kwargs['_test_ts']}.tar.zst"
-            outpath.write_bytes(b"\x28\xb5\x2f\xfd")  # zstd magic
-        from types import SimpleNamespace
-        return SimpleNamespace(returncode=0, stdout=b"", stderr=b"")
-
-    monkeypatch.setattr(backup_mod, "_run_subprocess", fake_run)
-
-    result = backup_mod.create_backup(
-        db_url="postgresql://x", images_root=tmp_path, tag="pre_migration",
-    )
-    assert "sql_path" in result
-    assert "tar_path" in result
-    assert result["tag"] == "pre_migration"
-    assert any("pg_dump" in c[0] for c in calls)
-    assert any(c[0] == "tar" for c in calls)
-
-
-@pytest.mark.asyncio
-async def test_find_latest_backup_by_tag(monkeypatch, tmp_path):
-    # Create two manifests with different tags.
-    backups_dir = tmp_path / "_backups"
-    backups_dir.mkdir()
-    import json
-    (backups_dir / "fc_20260101T000000Z.json").write_text(json.dumps({
-        "backup_id": "20260101T000000Z", "tag": "manual",
-        "created_at": "2026-01-01T00:00:00+00:00",
-        "sql_path": "/x/a.sql", "tar_path": "/x/a.tar.zst",
-    }))
-    (backups_dir / "fc_20260202T000000Z.json").write_text(json.dumps({
-        "backup_id": "20260202T000000Z", "tag": "pre_migration",
-        "created_at": "2026-02-02T00:00:00+00:00",
-        "sql_path": "/x/b.sql", "tar_path": "/x/b.tar.zst",
-    }))
-
-    found = backup_mod.find_latest_backup(tmp_path, tag="pre_migration")
-    assert found is not None
-    assert found["backup_id"] == "20260202T000000Z"
-
-    missing = backup_mod.find_latest_backup(tmp_path, tag="nonexistent")
-    assert missing is None

tests/test_task_run_signals.py

@@ -0,0 +1,258 @@
+"""FC-3i: Celery signal → task_run row tests.
+
+Uses task_always_eager so signals fire synchronously in-process during
+the test. Scoped autouse fixture keeps the eager mode confined to this
+file (other tests may break with eager mode enabled globally).
+
+Test assertions on signal-mutated rows use COLUMN SELECTS / db_sync.
+scalar (per reference_async_coredml_test_assertions). ORM attribute
+access on a row a signal handler just mutated risks MissingGreenlet
+across async/sync boundaries.
+"""
+
+import pytest
+from sqlalchemy import func, select
+
+from backend.app.celery_app import celery
+from backend.app.models import TaskRun
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture(autouse=True)
+def _eager_celery(monkeypatch):
+    """task_always_eager so signals fire in-process during the test."""
+    monkeypatch.setattr(celery.conf, "task_always_eager", True)
+    monkeypatch.setattr(celery.conf, "task_eager_propagates", False)
+
+
+@pytest.mark.asyncio
+async def test_prerun_inserts_running_row(db_sync):
+    """Fire a no-op task; assert task_run row inserted with running status."""
+    from backend.app.tasks.thumbnail import generate_thumbnail
+    generate_thumbnail.delay(999_999)  # nonexistent image_id → no-op
+
+    rows = db_sync.execute(
+        select(TaskRun.queue, TaskRun.target_id, TaskRun.task_name)
+        .where(TaskRun.target_id == 999_999)
+    ).all()
+    assert len(rows) == 1
+    queue, target_id, task_name = rows[0]
+    assert queue == "thumbnail"
+    assert target_id == 999_999
+    assert task_name.endswith(".generate_thumbnail")
+
+
+@pytest.mark.asyncio
+async def test_postrun_marks_ok_with_duration(db_sync):
+    from backend.app.tasks.thumbnail import generate_thumbnail
+    generate_thumbnail.delay(999_998)
+
+    row = db_sync.execute(
+        select(
+            TaskRun.status, TaskRun.finished_at, TaskRun.duration_ms,
+        ).where(TaskRun.target_id == 999_998)
+    ).one()
+    status, finished_at, duration_ms = row
+    assert status == "ok"
+    assert finished_at is not None
+    assert duration_ms is not None and duration_ms >= 0
+
+
+@pytest.mark.asyncio
+async def test_failure_records_error_type_and_message(db_sync, monkeypatch):
+    """Force a tracked task to raise; assert error fields populated."""
+    from backend.app.tasks.thumbnail import generate_thumbnail
+
+    def _explode(*a, **k):
+        raise ValueError("synthetic failure")
+    monkeypatch.setattr(
+        "backend.app.tasks.thumbnail.generate_thumbnail.run",
+        _explode,
+    )
+    with pytest.raises(ValueError):
+        generate_thumbnail.delay(999_997).get(propagate=True)
+
+    row = db_sync.execute(
+        select(
+            TaskRun.status, TaskRun.error_type, TaskRun.error_message,
+            TaskRun.finished_at,
+        ).where(TaskRun.target_id == 999_997)
+        .order_by(TaskRun.id.desc()).limit(1)
+    ).one()
+    status, error_type, error_message, finished_at = row
+    assert status == "error"
+    assert error_type == "ValueError"
+    assert "synthetic failure" in (error_message or "")
+    assert finished_at is not None
+
+
+@pytest.mark.asyncio
+async def test_soft_time_limit_records_timeout_status(db_sync, monkeypatch):
+    from celery.exceptions import SoftTimeLimitExceeded
+
+    from backend.app.tasks.thumbnail import generate_thumbnail
+
+    def _timeout(*a, **k):
+        raise SoftTimeLimitExceeded("simulated soft limit")
+    monkeypatch.setattr(
+        "backend.app.tasks.thumbnail.generate_thumbnail.run",
+        _timeout,
+    )
+    with pytest.raises(SoftTimeLimitExceeded):
+        generate_thumbnail.delay(999_996).get(propagate=True)
+
+    status = db_sync.execute(
+        select(TaskRun.status).where(TaskRun.target_id == 999_996)
+        .order_by(TaskRun.id.desc()).limit(1)
+    ).scalar_one()
+    assert status == "timeout"
+
+
+@pytest.mark.asyncio
+async def test_untracked_celery_internal_tasks_skip(db_sync):
+    """celery.chord_unlock is in the untracked list; firing it should
+    NOT insert a task_run row."""
+    before = db_sync.execute(select(func.count(TaskRun.id))).scalar_one()
+    # Send a chord_unlock-style task via apply (no real task to dispatch).
+    # The signal handler checks task.name against _UNTRACKED_TASK_NAMES;
+    # if name is in the list, no row is inserted. We verify by firing
+    # a regular task and asserting only ONE row appears, not two.
+    from backend.app.tasks.thumbnail import generate_thumbnail
+    generate_thumbnail.delay(999_995)
+    after = db_sync.execute(select(func.count(TaskRun.id))).scalar_one()
+    assert after - before == 1
+
+
+@pytest.mark.asyncio
+async def test_args_summary_truncated_to_255(db_sync, monkeypatch):
+    """Task with a very large repr(args); assert args_summary <= 255."""
+    from backend.app.tasks.thumbnail import generate_thumbnail
+
+    # Force a giant arg by patching repr(args) target — actually just
+    # call with a large int that produces a long repr.
+    big_id = int("9" * 200)  # 200-char repr
+    # The signal handler reads repr(args), so this naturally produces
+    # a long args_summary that should get truncated to 255.
+    try:
+        generate_thumbnail.delay(big_id).get(timeout=5, propagate=False)
+    except Exception:
+        pass
+
+    args_summary = db_sync.execute(
+        select(TaskRun.args_summary)
+        .where(TaskRun.task_name.endswith(".generate_thumbnail"))
+        .order_by(TaskRun.id.desc()).limit(1)
+    ).scalar_one()
+    assert args_summary is None or len(args_summary) <= 255
+
+
+@pytest.mark.asyncio
+async def test_error_message_truncated_to_2000(db_sync, monkeypatch):
+    from backend.app.tasks.thumbnail import generate_thumbnail
+
+    huge_message = "X" * 5000
+
+    def _raise_huge(*a, **k):
+        raise RuntimeError(huge_message)
+    monkeypatch.setattr(
+        "backend.app.tasks.thumbnail.generate_thumbnail.run", _raise_huge,
+    )
+    with pytest.raises(RuntimeError):
+        generate_thumbnail.delay(999_994).get(propagate=True)
+
+    error_message = db_sync.execute(
+        select(TaskRun.error_message).where(TaskRun.target_id == 999_994)
+        .order_by(TaskRun.id.desc()).limit(1)
+    ).scalar_one()
+    assert error_message is not None
+    assert len(error_message) <= 2000
+
+
+@pytest.mark.asyncio
+async def test_signal_handler_failure_does_not_break_task(db_sync, monkeypatch):
+    """Monkeypatch sync_session_factory in the signals module to raise;
+    assert the task still returns its result and the worker doesn't
+    propagate the signal-handler exception."""
+    from backend.app.tasks.thumbnail import generate_thumbnail
+
+    def _broken_factory():
+        raise RuntimeError("db down")
+
+    monkeypatch.setattr(
+        "backend.app.celery_signals.sync_session_factory", _broken_factory,
+    )
+
+    # Task should still execute (no-op on missing image) without raising.
+    result = generate_thumbnail.delay(999_993).get(timeout=5)
+    assert result is not None  # task ran; signal handler errors swallowed
+
+
+@pytest.mark.parametrize("args,expected", [
+    ((42,), 42),
+    ((42, "ignored"), 42),
+    (("123",), 123),
+    (("not-int",), None),
+    ((), None),
+    (None, None),
+])
+def test_target_id_extracted_from_first_int_arg(args, expected):
+    """Unit test the helper directly; no DB needed."""
+    from backend.app.celery_signals import _target_id_from_args
+    assert _target_id_from_args(args) == expected
+
+
+def test_finalize_sets_retry_status_with_retry_count(db_sync):
+    """Direct unit-shape test of _finalize's retry path. We can't
+    reliably trigger Celery's task_retry signal under eager mode with
+    monkeypatched .run (the bind=True 'self' injection breaks), so
+    test the finalize helper directly — that's the code the retry
+    signal handler delegates to anyway."""
+    from datetime import UTC, datetime
+
+    from backend.app.celery_signals import _finalize
+    from backend.app.models import TaskRun
+
+    # Seed a 'running' row that _finalize can target.
+    row = TaskRun(
+        celery_task_id="retry-test-tid",
+        queue="ml",
+        task_name="backend.app.tasks.fake.t",
+        target_id=999_992,
+        started_at=datetime.now(UTC),
+        status="running",
+    )
+    db_sync.add(row)
+    db_sync.commit()
+
+    _finalize(
+        "retry-test-tid", status="retry",
+        error_type="RuntimeError",
+        error_message="first attempt",
+        retry_count=1,
+    )
+
+    # Column-select assertions — _finalize mutated the row via its own
+    # session; re-read fresh state through db_sync.
+    db_sync.expire_all()
+    status, retry_count, error_type = db_sync.execute(
+        select(TaskRun.status, TaskRun.retry_count, TaskRun.error_type)
+        .where(TaskRun.celery_task_id == "retry-test-tid")
+    ).one()
+    assert status == "retry"
+    assert retry_count == 1
+    assert error_type == "RuntimeError"
+
+
+@pytest.mark.asyncio
+async def test_worker_hostname_recorded(db_sync):
+    """Sender's hostname should land in worker_hostname."""
+    from backend.app.tasks.thumbnail import generate_thumbnail
+    generate_thumbnail.delay(999_991)
+
+    hostname = db_sync.execute(
+        select(TaskRun.worker_hostname).where(TaskRun.target_id == 999_991)
+    ).scalar_one()
+    # Under eager mode hostname may be None or "celery@local"; just
+    # assert the column is touched (either string or None — never throws).
+    assert hostname is None or isinstance(hostname, str)

tests/test_tasks_admin.py

@@ -0,0 +1,158 @@
+"""FC-3k: admin Celery task integration tests.
+
+task_always_eager + signal handlers from FC-3i populate task_run.
+We assert the wrapper passes args through correctly and that
+task_run lifecycle status flips as expected.
+"""
+import pytest
+from sqlalchemy import func, select
+
+import backend.app.tasks.admin  # noqa: F401 — register tasks
+from backend.app.celery_app import celery
+from backend.app.models import Artist, ImageRecord, TaskRun
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture(autouse=True)
+def _eager_celery(monkeypatch):
+    monkeypatch.setattr(celery.conf, "task_always_eager", True)
+    monkeypatch.setattr(celery.conf, "task_eager_propagates", False)
+
+
+@pytest.fixture(autouse=True)
+def fake_images_root(monkeypatch, tmp_path):
+    monkeypatch.setattr("backend.app.tasks.admin.IMAGES_ROOT", tmp_path)
+
+
+# --- registration ----------------------------------------------------
+
+
+def test_delete_artist_cascade_task_registered():
+    assert (
+        "backend.app.tasks.admin.delete_artist_cascade_task"
+        in celery.tasks
+    )
+
+
+def test_bulk_delete_images_task_registered():
+    assert (
+        "backend.app.tasks.admin.bulk_delete_images_task"
+        in celery.tasks
+    )
+
+
+# --- delete_artist_cascade_task -------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_delete_artist_cascade_task_removes_artist_and_records_ok(
+    db_sync, tmp_path,
+):
+    from backend.app.tasks.admin import delete_artist_cascade_task
+
+    a = Artist(name="Doomed", slug="doomed")
+    db_sync.add(a)
+    db_sync.flush()
+    for i in range(2):
+        f = tmp_path / f"d{i}.jpg"
+        f.write_bytes(b"x")
+        db_sync.add(ImageRecord(
+            artist_id=a.id, path=str(f),
+            sha256=f"{i:064x}", size_bytes=10, mime="image/jpeg",
+            origin="imported_filesystem",
+        ))
+    db_sync.commit()
+    artist_id = a.id
+
+    result = delete_artist_cascade_task.delay(artist_id=artist_id).get()
+    assert result["summary"]["images_deleted"] == 2
+
+    surviving = db_sync.execute(
+        select(func.count(Artist.id)).where(Artist.id == artist_id)
+    ).scalar_one()
+    assert surviving == 0
+
+    # FC-3i task_run lifecycle: task should have an 'ok' row.
+    status = db_sync.execute(
+        select(TaskRun.status)
+        .where(TaskRun.task_name.endswith(".delete_artist_cascade_task"))
+        .order_by(TaskRun.id.desc()).limit(1)
+    ).scalar_one()
+    assert status == "ok"
+
+
+@pytest.mark.asyncio
+async def test_delete_artist_cascade_task_records_failure(
+    db_sync, monkeypatch,
+):
+    from backend.app.tasks.admin import delete_artist_cascade_task
+
+    def _boom(*a, **kw):
+        raise RuntimeError("synthetic cascade fail")
+    monkeypatch.setattr(
+        "backend.app.services.cleanup_service.delete_artist_cascade", _boom,
+    )
+    with pytest.raises(RuntimeError):
+        delete_artist_cascade_task.delay(artist_id=1).get()
+
+    row = db_sync.execute(
+        select(TaskRun.status, TaskRun.error_message)
+        .where(TaskRun.task_name.endswith(".delete_artist_cascade_task"))
+        .order_by(TaskRun.id.desc()).limit(1)
+    ).one()
+    assert row.status == "error"
+    assert "synthetic" in (row.error_message or "")
+
+
+# --- bulk_delete_images_task ----------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_images_task_removes_listed_images(
+    db_sync, tmp_path,
+):
+    from backend.app.tasks.admin import bulk_delete_images_task
+
+    a = Artist(name="B", slug="b")
+    db_sync.add(a)
+    db_sync.flush()
+    ids = []
+    for i in range(3):
+        f = tmp_path / f"b{i}.jpg"
+        f.write_bytes(b"x")
+        img = ImageRecord(
+            artist_id=a.id, path=str(f),
+            sha256=f"a{i:063x}", size_bytes=10, mime="image/jpeg",
+            origin="imported_filesystem",
+        )
+        db_sync.add(img)
+        db_sync.flush()
+        ids.append(img.id)
+    db_sync.commit()
+
+    result = bulk_delete_images_task.delay(image_ids=ids).get()
+    assert result["images_deleted"] == 3
+
+    surviving = db_sync.execute(
+        select(func.count(ImageRecord.id))
+        .where(ImageRecord.id.in_(ids))
+    ).scalar_one()
+    assert surviving == 0
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_images_task_idempotent_on_empty_list(db_sync):
+    from backend.app.tasks.admin import bulk_delete_images_task
+    result = bulk_delete_images_task.delay(image_ids=[]).get()
+    assert result["images_deleted"] == 0
+
+
+@pytest.mark.asyncio
+async def test_bulk_delete_images_task_reports_missing_ids(db_sync):
+    from backend.app.tasks.admin import bulk_delete_images_task
+    result = bulk_delete_images_task.delay(
+        image_ids=[9_999_998, 9_999_999],
+    ).get()
+    assert result["images_deleted"] == 0
+    assert sorted(result["missing_ids"]) == [9_999_998, 9_999_999]

tests/test_tasks_backup.py

@@ -0,0 +1,281 @@
+"""FC-3h: backup/restore Celery task integration tests.
+
+Uses task_always_eager for synchronous in-test execution.
+Subprocess + IMAGES_ROOT are monkeypatched so tests don't touch
+real /images or shell out to pg_dump.
+"""
+from datetime import UTC, datetime, timedelta
+from pathlib import Path
+
+import pytest
+from sqlalchemy import select
+
+from backend.app.celery_app import celery
+from backend.app.models import BackupRun, ImportSettings
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture(autouse=True)
+def _eager_celery(monkeypatch):
+    monkeypatch.setattr(celery.conf, "task_always_eager", True)
+    monkeypatch.setattr(celery.conf, "task_eager_propagates", False)
+
+
+@pytest.fixture(autouse=True)
+def fake_subprocess_and_images_root(monkeypatch, tmp_path):
+    monkeypatch.setattr("backend.app.tasks.backup.IMAGES_ROOT", tmp_path)
+    monkeypatch.setattr(
+        "backend.app.services.backup_service._DB_SUBPROCESS_TIMEOUT_S", 5,
+    )
+
+    class _FakeProc:
+        returncode = 0
+        stdout = b""
+        stderr = b""
+
+    def _fake_run(cmd, **kwargs):
+        if cmd[0] == "pg_dump":
+            i = cmd.index("-f")
+            Path(cmd[i + 1]).write_bytes(b"-- fake pg_dump\n")
+        elif cmd[0] == "tar" and "-cf" in cmd:
+            i = cmd.index("-cf")
+            Path(cmd[i + 1]).write_bytes(b"fake tar payload")
+        return _FakeProc()
+
+    monkeypatch.setattr("subprocess.run", _fake_run)
+
+
+def _seed_backup(db_sync, *, kind, status, started_at, tag=None,
+                 finished_at=None):
+    row = BackupRun(
+        kind=kind, status=status, tag=tag,
+        triggered_by="manual", started_at=started_at,
+        finished_at=finished_at or (started_at + timedelta(seconds=10)),
+        sql_path="/tmp/fake.sql" if kind == "db" else None,
+        tar_path="/tmp/fake.tar.zst" if kind == "images" else None,
+        manifest={},
+    )
+    db_sync.add(row)
+    db_sync.flush()
+    return row.id
+
+
+# --- backup_db_task --------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_backup_db_task_creates_backup_run_row_status_ok(db_sync):
+    from backend.app.tasks.backup import backup_db_task
+    result = backup_db_task.delay(tag=None, triggered_by="manual").get()
+    run_id = result["backup_run_id"]
+    row = db_sync.execute(
+        select(
+            BackupRun.kind, BackupRun.status, BackupRun.sql_path,
+            BackupRun.size_bytes, BackupRun.finished_at, BackupRun.error,
+        ).where(BackupRun.id == run_id)
+    ).one()
+    assert row.kind == "db"
+    assert row.status == "ok"
+    assert row.sql_path and row.sql_path.endswith(".sql")
+    assert row.size_bytes is not None and row.size_bytes > 0
+    assert row.finished_at is not None
+    assert row.error is None
+
+
+@pytest.mark.asyncio
+async def test_backup_db_task_records_failure_on_subprocess_error(db_sync, monkeypatch):
+    from backend.app.tasks.backup import backup_db_task
+
+    def _boom(*a, **kw):
+        raise RuntimeError("synthetic pg_dump fail")
+    monkeypatch.setattr("subprocess.run", _boom)
+
+    with pytest.raises(RuntimeError):
+        backup_db_task.delay().get()
+
+    row = db_sync.execute(
+        select(BackupRun.status, BackupRun.error)
+        .where(BackupRun.kind == "db")
+        .order_by(BackupRun.id.desc()).limit(1)
+    ).one()
+    assert row.status == "error"
+    assert "synthetic" in (row.error or "")
+
+
+@pytest.mark.asyncio
+async def test_backup_db_task_persists_tag(db_sync):
+    from backend.app.tasks.backup import backup_db_task
+    result = backup_db_task.delay(tag="pre-cutover").get()
+    tag = db_sync.execute(
+        select(BackupRun.tag).where(BackupRun.id == result["backup_run_id"])
+    ).scalar_one()
+    assert tag == "pre-cutover"
+
+
+# --- backup_images_task ---------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_backup_images_task_creates_backup_run(db_sync):
+    from backend.app.tasks.backup import backup_images_task
+    result = backup_images_task.delay().get()
+    row = db_sync.execute(
+        select(BackupRun.kind, BackupRun.status, BackupRun.tar_path)
+        .where(BackupRun.id == result["backup_run_id"])
+    ).one()
+    assert row.kind == "images"
+    assert row.status == "ok"
+    assert row.tar_path and row.tar_path.endswith(".tar.zst")
+
+
+# --- restore_db_task ------------------------------------------------
+
+
+@pytest.mark.asyncio
+async def test_restore_db_task_creates_restoring_marker_then_restored(db_sync):
+    from backend.app.tasks.backup import backup_db_task, restore_db_task
+    src = backup_db_task.delay().get()
+    src_id = src["backup_run_id"]
+
+    restore_db_task.delay(source_backup_run_id=src_id).get()
+
+    rows = db_sync.execute(
+        select(BackupRun.status, BackupRun.triggered_by)
+        .where(BackupRun.restored_from_id == src_id)
+    ).all()
+    assert len(rows) == 1
+    assert rows[0].status == "restored"
+    assert rows[0].triggered_by == "restore"
+
+
+@pytest.mark.asyncio
+async def test_restore_db_task_rejects_non_db_source(db_sync):
+    from backend.app.tasks.backup import restore_db_task
+
+    now = datetime.now(UTC)
+    img_id = _seed_backup(db_sync, kind="images", status="ok", started_at=now)
+    db_sync.commit()
+
+    with pytest.raises(ValueError):
+        restore_db_task.delay(source_backup_run_id=img_id).get()
+
+
+# --- prune_backups --------------------------------------------------
+
+
+def test_prune_backups_keeps_last_n_untagged_per_kind(db_sync):
+    from backend.app.tasks.backup import prune_backups
+
+    s = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    s.backup_db_keep_last_n = 2
+    s.backup_images_keep_last_n = 1
+    db_sync.commit()
+
+    now = datetime.now(UTC)
+    for i in range(4):
+        _seed_backup(db_sync, kind="db", status="ok",
+                     started_at=now - timedelta(hours=i))
+    for i in range(3):
+        _seed_backup(db_sync, kind="images", status="ok",
+                     started_at=now - timedelta(hours=i))
+    db_sync.commit()
+
+    result = prune_backups.apply().get()
+    assert result["db_deleted"] == 2
+    assert result["images_deleted"] == 2
+
+
+def test_prune_backups_protects_tagged_rows(db_sync):
+    from backend.app.tasks.backup import prune_backups
+
+    s = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    s.backup_db_keep_last_n = 1
+    db_sync.commit()
+
+    now = datetime.now(UTC)
+    _seed_backup(db_sync, kind="db", status="ok", started_at=now)
+    tagged_id = _seed_backup(
+        db_sync, kind="db", status="ok",
+        started_at=now - timedelta(days=30), tag="forever",
+    )
+    _seed_backup(db_sync, kind="db", status="ok",
+                 started_at=now - timedelta(days=1))
+    db_sync.commit()
+
+    prune_backups.apply().get()
+
+    surviving_tag = db_sync.execute(
+        select(BackupRun.tag).where(BackupRun.id == tagged_id)
+    ).scalar_one_or_none()
+    assert surviving_tag == "forever"
+
+
+def test_prune_backups_never_deletes_running_or_restoring(db_sync):
+    from backend.app.tasks.backup import prune_backups
+
+    s = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    s.backup_db_keep_last_n = 0
+    db_sync.commit()
+
+    ancient = datetime.now(UTC) - timedelta(days=30)
+    running_id = _seed_backup(db_sync, kind="db", status="running", started_at=ancient)
+    restoring_id = _seed_backup(db_sync, kind="db", status="restoring", started_at=ancient)
+    db_sync.commit()
+
+    prune_backups.apply().get()
+
+    statuses = db_sync.execute(
+        select(BackupRun.status).where(BackupRun.id.in_([running_id, restoring_id]))
+    ).scalars().all()
+    assert set(statuses) == {"running", "restoring"}
+
+
+# --- backup_db_nightly ----------------------------------------------
+
+
+def test_nightly_skips_when_disabled(db_sync):
+    from backend.app.tasks.backup import backup_db_nightly
+
+    s = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    s.backup_db_nightly_enabled = False
+    db_sync.commit()
+
+    result = backup_db_nightly.apply().get()
+    assert "skipped" in result and "disabled" in result["skipped"]
+
+
+def test_nightly_skips_when_hour_mismatch(db_sync):
+    from backend.app.tasks.backup import backup_db_nightly
+
+    s = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    s.backup_db_nightly_enabled = True
+    s.backup_db_nightly_hour_utc = (datetime.now(UTC).hour + 12) % 24
+    db_sync.commit()
+
+    result = backup_db_nightly.apply().get()
+    assert "skipped" in result and "hour=" in result["skipped"]
+
+
+def test_nightly_dispatches_when_enabled_at_configured_hour(db_sync):
+    from backend.app.tasks.backup import backup_db_nightly
+
+    s = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    s.backup_db_nightly_enabled = True
+    s.backup_db_nightly_hour_utc = datetime.now(UTC).hour
+    db_sync.commit()
+
+    result = backup_db_nightly.apply().get()
+    assert "dispatched" in result