feat(mcp): Steward's own backups: the runs, the next one, checking and making one (#5836)

list_backups, check_backup and make_backup are Settings → Backups for an
agent, admin only like the page. make_backup runs the page's job in the
background and is audited as "<user> via MCP". Restoring stays in the
browser. get_overview names a scheduled backup that failed, is two hours
past its slot, or has no bucket.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

steward/mcp/tools/overview.py

@@ -7,8 +7,10 @@ from datetime import timedelta
 from sqlalchemy import select
 
 from steward.alerts import service as alert_svc
+from steward.backup import job as backup_job
 from steward.core import alerts
 from steward.core import anomalies as anom
+from steward.core.settings import get_all_settings
 from steward.hosts import service as host_svc
 from steward.mcp._context import mcp_app
 from steward.models.ansible import AnsibleRun, AnsibleRunStatus
@@ -22,6 +24,9 @@ from ._shape import age_s, db_session, iso, now_utc, rnd
 logger = logging.getLogger(__name__)
 
 FINDINGS_CAP = 50
+# A scheduled backup this long past its slot with none made is overdue: one
+# 15-minute tick and the hour a failed try waits, with room to spare.
+BACKUP_OVERDUE_AFTER = timedelta(hours=2)
 _SEVERITY = {"critical": 0, "warning": 1}
 
 
@@ -202,6 +207,41 @@ def anomaly_findings(open_now) -> list[dict]:
     return out
 
 
+def backup_findings(settings: dict, running: bool, now) -> list[dict]:
+    """Steward's own backups (#5836), while scheduled ones are on: the bucket
+    isn't set up, the last one failed (since the first of the failures in a
+    row), or none has been made since the day's slot."""
+    from steward.core.formatting import zone
+    if running or not settings.get("backup.enabled"):
+        return []
+    subject, nxt = "Steward backups", "list_backups()"
+    if backup_job.target_from(settings) is None:
+        return [finding("warning", "backup_not_set_up", subject,
+                        "Scheduled backups are on, but the bucket isn't set up, so none are made",
+                        next=nxt)]
+    history = settings.get("backup.history") or []
+    last_ok = backup_job._moment(settings.get("backup.last_success_at"))
+    good = (f"the newest good one was sent {iso(last_ok)}" if last_ok
+            else "none has been sent yet")
+    failing = []
+    for entry in history:
+        if entry.get("ok"):
+            break
+        failing.append(entry)
+    if failing:
+        tries = f" ({len(failing)} tries in a row)" if len(failing) > 1 else ""
+        return [finding("warning", "backup_failed", subject,
+                        f"The last backup failed{tries}: {failing[0].get('error') or 'no reason recorded'}; {good}",
+                        since=backup_job._moment(failing[-1].get("at")), next=nxt)]
+    slot = backup_job.last_slot(now, int(settings.get("backup.schedule_hour", 3)),
+                                zone(settings.get("backup.schedule_tz")))
+    if history and (last_ok is None or last_ok < slot) and now - slot >= BACKUP_OVERDUE_AFTER:
+        return [finding("warning", "backup_overdue", subject,
+                        f"No backup has been made since the scheduled one at {iso(slot)}; {good}",
+                        since=slot, next=nxt)]
+    return []
+
+
 async def _plugin_findings(server) -> tuple[list[dict], list[str]]:
     found, unavailable = [], []
     for plugin, source in sorted((getattr(server, "overview_sources", None) or {}).items()):
@@ -225,7 +265,8 @@ async def get_overview() -> dict:
     agents that stopped reporting, hosts over their CPU, memory or disk
     limits (all cutoffs are set under Settings → Thresholds), Ansible runs that failed
     in the last 24h, metrics going on now well above what is normal for them
-    at this hour (metric anomalies), and anything enabled plugins add
+    at this hour (metric anomalies), Steward's own scheduled backup failing or
+    overdue, and anything enabled plugins add
     (unhealthy containers, for one). Each finding carries the ids to follow
     and `next`, the call that shows its evidence.
 
@@ -254,6 +295,7 @@ async def get_overview() -> dict:
         windows = await alert_svc.list_maintenance_windows(db)
         open_anomalies = await anom.open_anomalies(db)
         to_review = await anom.to_review_count(db)
+        settings = await get_all_settings(db)
 
     thresholds = app.config.get("THRESHOLDS") or {}
     listed, silent, speaking = fold_agent_silence(active, agent)
@@ -261,7 +303,8 @@ async def get_overview() -> dict:
                 + alert_findings(listed, silent)
                 + agent_findings(hosts, agent, thresholds, silent, speaking)
                 + run_findings(runs)
-                + anomaly_findings(open_anomalies))
+                + anomaly_findings(open_anomalies)
+                + backup_findings(settings, backup_job.is_running(), now))
     plugin_found, unavailable = await _plugin_findings(getattr(app, "mcp_server", None))
     findings = rank(findings + plugin_found)