Kordant/tasks/web-production/12-uptime-monitoring.md

12. Uptime & Performance Monitoring

meta: id: web-production-12 feature: web-production priority: P2 depends_on: [] tags: [observability, uptime, production]

objective:

• Monitor application uptime and performance from external vantage points to ensure reliability

deliverables:

• External uptime monitoring (Pingdom, UptimeRobot, or Datadog Synthetics)
• Synthetic monitoring for critical user journeys
• Performance budget enforcement
• Status page for incident communication

steps:

1. Set up uptime monitoring:
   • Configure checks for homepage, API health, dashboard
   • Check from multiple regions (US East, US West, EU)
   • 1-minute interval checks
   • Alert on 2 consecutive failures
2. Implement synthetic monitoring:
   • Signup flow: homepage → signup → verify email
   • Login flow: login → dashboard → view alerts
   • Billing flow: dashboard → pricing → checkout (test mode)
   • DarkWatch flow: dashboard → darkwatch → add watchlist item
3. Set performance budgets:
   • LCP (Largest Contentful Paint) < 2.5s mobile, < 1.5s desktop
   • FID (First Input Delay) < 100ms
   • CLS (Cumulative Layout Shift) < 0.1
   • TTFB (Time to First Byte) < 200ms
   • API response p95 < 200ms
4. Configure alerting:
   • Downtime alert via Slack/SMS
   • Performance degradation alert (LCP > 3s)
   • SSL certificate expiry alert (30 days before)
   • Domain expiry alert (30 days before)
5. Set up status page:
   • Use statuspage.io or instatus.com
   • Auto-update from monitoring checks
   • Subscribe users for incident notifications
   • Post incident updates and post-mortems

tests:

• Integration: Verify monitoring catches simulated outage
• Performance: Confirm synthetic tests complete successfully
• Alert: Test alert channels with deliberate failure

acceptance_criteria:

• Uptime monitoring checking every 60 seconds from 3+ regions
• 99.9% uptime SLA measured over 30 days
• Synthetic tests covering signup, login, and core flows
• Performance budget alerts for LCP > 2.5s
• Status page accessible and auto-updating
• SSL certificate expiry alert 30 days in advance

validation:

• Simulate outage → alert received within 2 minutes
• Check status page → shows incident with timeline
• Run synthetic test → completes in <30 seconds
• Lighthouse CI shows all metrics within budget

notes:

• UptimeRobot free tier: 50 monitors, 5-minute intervals
• Pingdom more reliable but paid
• Consider using Checkly for synthetic monitoring with JS