feat(docker): collect + persist /system/df image/disk usage (agent 1.6.0)

Backend for the image/disk panel (milestone 77 #942). Agent gains collect_disk_usage() — one /system/df call (gated on containers existing, so Docker-less hosts pay nothing), surfacing reclaimable bytes (image size held by unreferenced images), layers/containers/volumes/build-cache sizes, and the top 50 images by size. Emitted as sample["docker_disk"]; host_agent ingest tracks the newest sample's copy and hands it to the docker capability as a 5th arg. New current-state tables docker_disk_usage (one row/host) + docker_images (per-host image rows), docker_007 migration; ingest upserts the summary and replaces the image set per host (stale images pruned). Unit tests for the df parsing/reclaimable math + build_sample gating; integration test for persistence + image-set replacement. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016Jg27rgypiW2efULXJDtMC
2026-06-18 21:51:14 -04:00
parent faecac3ec6
commit a840d6f823
7 changed files with 367 additions and 3 deletions
@@ -0,0 +1,55 @@
+"""Docker disk usage + image storage tables
+
+Adds docker_disk_usage (one per-host /system/df summary row) and docker_images
+(the heavy-hitter image storage records). Both current-state, host-scoped,
+re-reported each interval. Additive create_table.
+
+Revision ID: docker_007_disk_usage
+Revises: docker_006_metric_rollup
+Create Date: 2026-06-19
+"""
+from typing import Sequence, Union
+from alembic import op
+import sqlalchemy as sa
+
+revision: str = "docker_007_disk_usage"
+down_revision: Union[str, None] = "docker_006_metric_rollup"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "docker_disk_usage",
+        sa.Column("host_id", sa.String(length=36), nullable=False),
+        sa.Column("layers_size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("images_size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("images_reclaimable", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("containers_size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("volumes_size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("build_cache_size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("images_total", sa.Integer(), nullable=False, server_default="0"),
+        sa.Column("images_active", sa.Integer(), nullable=False, server_default="0"),
+        sa.Column("containers_count", sa.Integer(), nullable=False, server_default="0"),
+        sa.Column("volumes_count", sa.Integer(), nullable=False, server_default="0"),
+        sa.Column("updated_at", sa.DateTime(timezone=True), nullable=False),
+        sa.ForeignKeyConstraint(["host_id"], ["hosts.id"], ondelete="CASCADE"),
+        sa.PrimaryKeyConstraint("host_id"),
+    )
+    op.create_table(
+        "docker_images",
+        sa.Column("host_id", sa.String(length=36), nullable=False),
+        sa.Column("image_id", sa.String(length=64), nullable=False),
+        sa.Column("repo_tag", sa.String(length=512), nullable=False, server_default="<none>"),
+        sa.Column("size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("shared_size", sa.BigInteger(), nullable=False, server_default="0"),
+        sa.Column("containers", sa.Integer(), nullable=False, server_default="0"),
+        sa.Column("updated_at", sa.DateTime(timezone=True), nullable=False),
+        sa.ForeignKeyConstraint(["host_id"], ["hosts.id"], ondelete="CASCADE"),
+        sa.PrimaryKeyConstraint("host_id", "image_id"),
+    )
+
+
+def downgrade() -> None:
+    op.drop_table("docker_images")
+    op.drop_table("docker_disk_usage")