Observability
Observability uses the three pillars—metrics, logs, and traces—to understand internal system state for troubleshooting and optimization.
Purpose
Infer internal system state from external outputs for fast troubleshooting and optimization.
Mechanism
Metrics (Prometheus): time-series values (CPU/latency/error rate). Logs (Loki/ELK): discrete event text. Traces (Jaeger): request paths across services. OpenTelemetry unifies collection.
Use cases
- System monitoring
- Troubleshooting
- Performance optimization