observability: scrape CloudNativePG postgres metrics (follow-up to #239) (#242)

Follow-up to #239. Every CNPG instance pod already exposes Prometheus metrics on `:9187` (`metrics` port), but nothing scraped them. Adds a single namespace-wide `VMPodScrape` (`namespaceSelector.any`, `selector cnpg.io/podRole=instance`) so the observability VMAgent collects postgres metrics for **all ~23 CNPG clusters** across the estate (authentik, grafana, woodpecker, artifactapi, puppet, litellm, …). No chart changes. Pairs well with the CNPG grafana.com dashboard (20417) as a further follow-up.

Reviewed-on: #242
Co-authored-by: Ben Vincent <ben@unkin.net>
Co-committed-by: Ben Vincent <ben@unkin.net>

apps/base/observability/kustomization.yaml

@@ -8,3 +8,4 @@ resources:
   - vmagent.yaml
   - gateway.yaml
   - httproute.yaml
+  - vmpodscrape-cnpg.yaml

apps/base/observability/vmpodscrape-cnpg.yaml

@@ -0,0 +1,17 @@
+---
+# Scrape metrics (:9187) from every CloudNativePG postgres instance pod in any
+# namespace. Picked up by the observability VMAgent (selectAllByDefault).
+apiVersion: operator.victoriametrics.com/v1beta1
+kind: VMPodScrape
+metadata:
+  name: cnpg-clusters
+  namespace: observability
+spec:
+  namespaceSelector:
+    any: true
+  selector:
+    matchLabels:
+      cnpg.io/podRole: instance
+  podMetricsEndpoints:
+    - port: metrics
+      path: /metrics