What is monitoring?
Monitoring means continuously collecting and tracking system data — servers, apps, databases, and APIs — to detect failures, slow APIs, memory leaks, and downtime.
What is observability?
Observability means understanding why a problem happened.
| Monitoring | Observability |
|---|---|
| Something is wrong | Why it is wrong |
| Dashboards & thresholds | Investigation with context |
| Known failure modes | Unknown / novel failures |
Three pillars of observability
| Pillar | Answers | Examples |
|---|---|---|
| Metrics | What is happening? | CPU %, request rate, error rate |
| Logs | What happened? | App errors, access logs |
| Traces | Where did time go? | Request path across services |
Car dashboard analogy
- Metrics — speedometer, fuel gauge
- Logs — trip computer / event history
- Traces — full route of a journey across roads
Why production monitoring matters
Without monitoring you discover outages from customers. With monitoring you catch:
- High CPU / memory before crash
- Rising 5xx rates
- Slow endpoints (p95 / p99)
- Disk filling up
Aim for metrics for symptoms, logs for detail, and traces for distributed latency — together they form observability.