[ci skip] fix post-NFS-migration issues: MySQL GR, Loki, grampsweb, alerts

- Loki: reduce memory limit from 6Gi to 4Gi (within LimitRange max)
- Grampsweb: increase memory to 2Gi (was OOMKilled at 512Mi)
- Fix PostgreSQLDown alert: check pod readiness instead of deployment
- Fix MySQLDown alert: check StatefulSet replicas instead of deployment
- Fix RedisDown alert: check StatefulSet replicas instead of deployment
- Fix NFSServerUnresponsive: aggregate all NFS versions cluster-wide
- Fix Uptime Kuma healthcheck: handle nested list heartbeat format
- Update etcd backup image to registry.k8s.io/etcd:3.6.5-0

stacks/platform/modules/monitoring/loki.yaml

@@ -70,10 +70,10 @@ singleBinary:
   resources:
     requests:
       cpu: 250m
-      memory: 4Gi
+      memory: 2Gi
     limits:
       cpu: "1"
-      memory: 6Gi
+      memory: 4Gi
 
 # Zero out replica counts of other deployment modes
 backend: