Skip to content

Latest commit

 

History

History
40 lines (29 loc) · 1.43 KB

File metadata and controls

40 lines (29 loc) · 1.43 KB

ADR-005: Observability Stack

Status

Accepted

Context

The IDP needs full-stack observability covering metrics, traces, and alerting. The solution must be CNCF-native, open source, and integrate with our existing ArgoCD GitOps workflow.

Decision

Use the CNCF-native observability stack:

  • Prometheus (metrics collection and alerting)
  • Grafana 12.x (visualization and dashboards)
  • OpenTelemetry Collector (telemetry pipeline for app instrumentation)

Deploy Prometheus and Grafana via kube-prometheus-stack Helm chart, which bundles node-exporter, kube-state-metrics, and Grafana with pre-configured datasources. OTel Collector runs as a DaemonSet receiving OTLP from applications and exporting to Prometheus via remote write.

Alternatives Considered

Tool Reason for Rejection
Datadog Commercial, vendor lock-in, cost
ELK Stack Heavier footprint, log-focused not metrics-first
CloudWatch AWS-only, limited dashboard flexibility
Mimir/VictoriaMetrics Solve scale problems we don't have in a demo

Consequences

  • Multiple components to wire (Prometheus, Grafana, OTel, exporters)
  • OTel Collector config requires manual verification of endpoints
  • Prometheus remote write receiver must be explicitly enabled
  • kube-prometheus-stack provides batteries-included setup (reduces config)
  • Sample app uses OTel auto-instrumentation for zero-code traces/metrics