Observability Pack (Prometheus + Grafana + OpenTelemetry)

Overview: Vendor-neutral observability that brings metrics, logs, and traces together with opinionated dashboards.
Key Capabilities: OTel SDK presets for Python/Node/Go; Prometheus scrape configs; Loki pipelines; optional Tempo for traces;
Grafana dashboards for API health, latency SLOs, error budgets, and business KPIs; alerting rules and on-call hooks.
Architecture: OTel exporters and service discovery; Prometheus TSDB retention; Loki compactor; curated Grafana JSON with drill-downs.
Security & Compliance: TLS for agents/gateways; auth on dashboards; multi-tenant folders; PII-safe logging and retention guidance.
Performance & Ops: Cardinality management; exemplars linking traces↔metrics; HA topologies (Thanos/Cortex) and sizing calculator.
Quick Start: docker compose -f observability.yml up → import dashboards → set labels/exemplars → configure alert receivers.
Deliverables: Configs, dashboards JSON, alert rules, docs, and SLO governance runbooks.
FAQ: Clean mapping notes for Datadog/New Relic; tracing is optional but recommended.