NIP: dependency slow (SSM / CBS)

Alert: NipDependencySlow · Severity: Warning · Requires app /metrics

What this means

p95 latency for an external dependency (ssm or cbs) exceeded 5 seconds for 5+ minutes.

Business impact

• Transfers slow; may timeout and fail with code 96 or wedge workers.
• Workbook game-day test: SSM slowdown should show on this panel within 10 seconds.

Confirm

histogram_quantile(0.95, sum by (dependency, le) (rate(nip_dependency_seconds_bucket[5m])))

Check which dependency label fired: ssm vs cbs.

Fix

Dependency	Actions
SSM	AWS status, IAM/credentials, region; check app logs for decrypt errors
CBS	Core banking team; network to CBS endpoint; recent deploy

1. Correlate with transfer failure spike.
2. If workers maxed → all workers busy or restart web per health probe.

Escalate if

p95 >30s or failure rate spikes simultaneously.