Incident Replay From Logs

Some incidents are reproducible. The replay technique that makes fix-validation reliable.

Capture inputs

From logs: identify the request sequence that caused the failure.

Strip PII; preserve the failure pattern.

Replay

Run the captured inputs through pre-prod. Verify the failure reproduces.

Apply the proposed fix; verify it now passes.

Limits

Stateful systems are hard to replay.

Some failures depend on timing that's hard to recreate.