Why should you care?

Deploy success does not mean users succeed — prod monitoring catches runtime issues.

See it live — copy this example

Paste into your MeanVerse project (Angular + Express + MongoDB), then run with ng serve / node / mongosh as noted.

# Alert rule (Prometheus)
- alert: MeanVerseTransferErrorRate
  expr: rate(http_requests_total{route="/api/transfers",status=~"5.."}[5m]) / rate(http_requests_total{route="/api/transfers"}[5m]) > 0.02
  for: 5m
  labels: { severity: page }
  annotations: { summary: "Transfer error rate above 2%" }

# Runbook link in annotation

What happened?

• PromQL compares 5xx rate to total transfer requests.
• fires 5 minutes sustained to reduce noise.
• Page severity wakes on-call with runbook.

Practice next

1. Define SLIs: availability, latency, error rate per endpoint.
2. Connect alerts to PagerDuty/Opsgenie.
3. Write runbook: check Mongo, recent deploy, dependency status.
4. Add synthetic transfer every 5 min from canary region.
5. Dashboard per tenant tier for enterprise customers.

Remember

Prod monitoring = SLOs + alerts + runbooks. Error budget guides release pace. Post-incident reviews improve system.

2am transfer failures

Mongo primary step-down causes brief 503s.

Outcome: Alert pages on-call; runbook confirms failover; auto-recovery in 4 min logged.