Log Analysis Guide

Correlation-ID-based RCA and operational KPIs.

Last updated: 2026-05-27 Section: Troubleshooting

Effective RCA rebuilds the transaction chain with Correlation ID, not isolated log lines.

Analysis workflow

Confirm incident window and affected APIs
Collect gateway → upstream logs by Correlation ID
Classify error codes, latency, and retry patterns
Correlate DB/HTTP/broker metrics
Apply temporary and permanent fixes separately

Key KPIs

KPI	Description	Target example
MTTA	Time to detect	< 5 min
MTTR	Time to recover	< 30 min
P95 Latency	95th percentile delay	Within SLO
Error Rate	Failure ratio	< 1%

{
  "timestamp": "...",
  "level": "ERROR",
  "service": "gateway-payment",
  "correlationId": "trace-uuid",
  "routeId": "payment-v2",
  "upstream": "svc-billing",
  "errorCode": "UPSTREAM_TIMEOUT",
  "latencyMs": 742
}