Monitoring & Alerting

Executive Summary

In brief: Outage detection and log review.

Why it matters: Early warning when portal is down.

Cost: €0–20/month for external tools optional.

Status: Basic checks; alerting partial.

Decision needed: Decide on paid uptime alerting if needed

Executive Overview — all summaries for decision-makers.

Status: Not yet configured (manual setup in external services)

Recommended Stack (cost-effective)

Purpose	Tool	Cost
Uptime	Better Stack or UptimeRobot	Free tier
Errors	Sentry	Free tier
Logs	Render Dashboard	included
Cron	Render Cron Job	included

Uptime Checks

URL	Interval	Alert
https://api.mqa.group/health	1 min	Email + Slack
https://portal.mqa.group/login	5 min	Email
https://api.mqa.group/api/auth/status	5 min	Email

Header for auth/status: not required (public endpoint).

Render Alerts

Dashboard → Service → Notifications:

• Deploy failed
• Instance unhealthy
• OOM / restart loop

Sentry Integration (TODO)

Backend

npm install @sentry/node --save

Env: SENTRY_DSN (Render secret)

Frontend

npm install @sentry/react --save

Env: VITE_SENTRY_DSN (DSN only, not a secret)

Log Retention

Render Free/Standard: limited retention. For audit:

• Important security events already in DB (AuditLog)
• Optional monthly export for compliance archive

On-Call (proposal)

Severity	Response time	Channel
Portal down	1h	Phone + email
Degraded (slow)	4h	Email
Security incident	immediate	Lead Dev + MQA IT

Runbook: BACKUP-RESTORE-RUNBOOK

Metrics to Watch

Render Dashboard → Metrics:

• CPU / memory backend
• HTTP latency p95
• Postgres connections

For AI features: Anthropic/OpenRouter usage dashboard separately.