# Monitoring and Observability
Confidence: high
Last verified: 2026-05-22
Generation: human_only


## TL;DR

Monitoring tells you what's wrong; observability lets you ask why. Three pillars: metrics (numeric time-series, counters/gauges), logs (timestamped event records), traces (end-to-end request flow across services). The 'Golden Signals': latency, traffic, errors, saturation.

## Core Explanation

Tools: Prometheus (metrics, pull-based), Grafana (dashboards), ELK/OpenSearch (logs), Jaeger/Zipkin (traces), OpenTelemetry (vendor-neutral instrumentation). SLOs (Service Level Objectives) define acceptable reliability: e.g., 99.9% availability with <100ms p99 latency. Alerting: alert on symptoms (user-facing impact), not causes. RED method: Rate, Errors, Duration per service.

## Further Reading

- [Google SRE Book](https://sre.google/books/)