feat(alerts): cover OOM kills, restart loops and evictions

The OOMKilled container behind the immich crash loop was invisible: PodCrashLooping only fires once kubelet has already given up and started the backoff.

prometheus-stack/k8s/alerts.yaml

@@ -27,6 +27,33 @@ spec:
             summary: "Pod is crash looping"
             description: "Container {{ $labels.container }} in {{ $labels.namespace }}/{{ $labels.pod }} is in CrashLoopBackOff."
 
+        - alert: ContainerOOMKilled
+          expr: max_over_time(kube_pod_container_status_terminated_reason{reason="OOMKilled"}[15m]) >= 1
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Container was OOMKilled"
+            description: "Container {{ $labels.container }} in {{ $labels.namespace }}/{{ $labels.pod }} was killed for exceeding its memory limit. Raise the limit or reduce the workload."
+
+        - alert: ContainerRestartingTooOften
+          expr: max by (namespace, pod, container) (increase(kube_pod_container_status_restarts_total[30m])) > 3
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Container restarting too often"
+            description: "Container {{ $labels.container }} in {{ $labels.namespace }}/{{ $labels.pod }} restarted {{ $value }} times in the last 30 minutes."
+
+        - alert: PodEvicted
+          expr: max_over_time(kube_pod_status_reason{reason="Evicted"}[15m]) >= 1
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Pod was evicted"
+            description: "Pod {{ $labels.namespace }}/{{ $labels.pod }} was evicted, usually for node disk or memory pressure."
+
         - alert: PersistentVolumeClaimFillingUp
           expr: kubelet_volume_stats_available_bytes / kubelet_volume_stats_capacity_bytes < 0.15
           for: 15m