Fix observability issues and refactor metrics stack to ArgoCD

- Resolved "connection refused" on startup probe by explicitly mapping management.server.port in application.yaml.
- Adjusted startupProbe to a 70s window (10s delay + 12 * 5s) to accommodate application initialization.
- Externalized VictoriaMetrics stack (Operator, Single, Agent, Alert) into separate ArgoCD Application manifests.
- Removed bundled VictoriaMetrics templates from the MovieNight Helm chart.
- Configured vmagent with Kubernetes pod-based service discovery for metrics scraping.
- Ensured consistency between Helm values and externalized observability components.

Co-authored-by: devitq <118541411+devitq@users.noreply.github.com>

deploy/argocd/victoriametrics-operator.yaml

@@ -0,0 +1,32 @@
+apiVersion: argoproj.io/v1alpha1
+kind: Application
+metadata:
+  name: victoriametrics-operator
+  namespace: argocd
+spec:
+  project: default
+  source:
+    repoURL: https://victoriametrics.github.io/helm-charts/
+    chart: victoria-metrics-operator
+    targetRevision: 0.38.0
+    helm:
+      releaseName: victoriametrics-operator
+      valuesObject:
+        admissionWebhooks:
+          enabled: false
+        createCRD: true
+        operator:
+          # This enables the operator to watch for CRs in all namespaces
+          # or we can specify namespaces.
+          disableNamespaceRestriction: true
+  destination:
+    server: https://kubernetes.default.svc
+    namespace: observability
+  syncPolicy:
+    automated:
+      prune: true
+      selfHeal: true
+    syncOptions:
+      - CreateNamespace=true
+      - ApplyOutOfSyncOnly=true
+      - ServerSideApply=true