Opservabilnost je sposobnost razumijevanja unutarnjeg stanja sustava iz njegovih vanjskih izlaza — kroz logove, metrike i tragove. Ona je neophodno potrebna za upravljanje, otklanjanje grešaka i održavanje sustava (pogotovo distribuiranih), gdje ne možete upravljati onim što ne vidite.
Tri stupa opservabilnosti
LOGS → timestamped records of events (what happened) → detailed, for debugging specific issues
METRICS → numerical measurements over time (CPU, latency, request rate, error rate) →
aggregate health/performance; dashboards; alerting
TRACES → follow a request's path through the system (across services) → understand flows,
find bottlenecks/failures in DISTRIBUTED systems (which service was slow?)
→ together: understand WHAT happened, the OVERALL state, and the PATH of requests.
