ragleap-observability v0.2.0 release notes

Added

Prometheus server and Grafana, live-verified end-to-end. Deployment/Service/PVC templates for Prometheus, plus Grafana (Deployment, Service, Secret, datasource ConfigMap).
Loki deployment and Promtail DaemonSet for log shipping, live-verified end-to-end: 21+ real log streams confirmed with correct namespace/pod/container labels.
Grafana datasource ConfigMap conditionally wires in Loki alongside Prometheus.
New loki: and promtail: blocks in values.yaml.

postgres-exporter runAsNonRoot fix (runAsUser: 65534) for named-user images.
Cross-namespace DNS lookup failure in Promtail's Loki client — fixed by using the fully-qualified service name.
kubernetes_sd_configs produced 0 active targets: fixed a node-scoping HOSTNAME issue and replaced a custom regex with Grafana's canonical __path__ pattern, plus added the cri pipeline stage.
A missing newline had silently corrupted a ConfigMap's YAML document boundary, causing a cascade of confusing symptoms across several helm upgrade cycles.
Raised Loki's ingestion rate limit to handle replay bursts from Promtail restarts.
Removed the temporary static-pod-logs diagnostic job, which was silently starving the real job of every file via shared position tracking.

Full pipeline confirmed live, layer by layer (Prometheus targets, Grafana health/datasources, Loki series query returning 21 real labeled streams, Promtail /ready returning Ready).

Neo4j Prometheus support unverified; neo4jExporter stays disabled.
AlertManager, SLO/SLI dashboards, recurring log-shipping health check remain unbuilt.
Loki retention (7 days) unverified for real sizing needs.

Full details: CHANGELOG.md