feat(db): 控制面 PG 定时备份、迁移前快照与原子恢复

deploy/alerts/felis-alerts.yaml

@@ -5,6 +5,7 @@
 # felis_* series come from two processes:
 #   - felis-operator pod :8080/metrics      → felis_servers_total, felis_start_duration_seconds
 #   - felis-api internal :8081/metrics      → felis_image_build_failures_total
+#   - node-exporter textfile collector      → felis_db_backup_* (felis-db-backup.timer)
 # node_* / kube_* series come from node-exporter / kube-state-metrics.
 groups:
   - name: felis.rules
@@ -67,3 +68,29 @@ groups:
           description: >-
             PostgreSQL, the control plane, the registry and game servers share one
             node; sustained memory pressure risks OOM kills.
+  - name: felis.backup.rules
+    rules:
+      - alert: FelisDBBackupStale
+        expr: time() - max(felis_db_backup_last_success_timestamp_seconds) > 26 * 3600
+        for: 10m
+        labels:
+          severity: critical
+        annotations:
+          summary: "no control-plane database backup in over 26h"
+          description: >-
+            felis-db-backup.timer runs daily; the newest bundle is more than a day
+            old. Read `journalctl -u felis-db-backup` on the host, then take one now
+            with `sudo felis db backup` (troubleshooting §16).
+      - alert: FelisDBBackupMetricMissing
+        expr: absent(felis_db_backup_last_success_timestamp_seconds)
+        for: 2h
+        labels:
+          severity: warning
+        annotations:
+          summary: "database backup freshness is not being scraped"
+          description: >-
+            No felis_db_backup_last_success_timestamp_seconds series, so
+            FelisDBBackupStale cannot fire. Point node-exporter's
+            --collector.textfile.directory at the directory of
+            FELIS_DB_BACKUP_METRICS (default /var/lib/node_exporter/textfile_collector)
+            (troubleshooting §16).