fix(observability): 🔇 silence managed-K8s false alerts + bump backup deadline to 4h

- Disable kubernetesSystemControllerManager, kubeScheduler, kubernetesSystemScheduler
  alert rules on dev, benchmark, edp clusters (unreachable on managed K8s)
- Bump forgejo s3 backup activeDeadlineSeconds 7200→14400 (2h→4h) across
  all instances; deadline hit Jun 20-21 on heavy sync

otc/edp.buildth.ing/stacks/observability-client/vm-client-stack/values.yaml

@@ -201,13 +201,13 @@ defaultRules:
       create: true
       rules: {}
     kubernetesSystemControllerManager:
-      create: true
+      create: false
       rules: {}
     kubeScheduler:
-      create: true
+      create: false
       rules: {}
     kubernetesSystemScheduler:
-      create: true
+      create: false
       rules: {}
     kubeStateMetrics:
       create: true