Skip to content

Monitoring

Purpose

What observability actually exists in this repository today - and what doesn't, so nobody goes looking for a dashboard that was never built.

What exists

  • Cloudflare dashboard - each Worker's built-in request/error metrics (Workers & Pages → select Worker → Metrics). No code in this repository configures this; it's a standard Cloudflare account feature, available to whoever has dashboard access.
  • wrangler tail - live log streaming per Worker, see Logs.
  • Production smoke test - npm run smoke:production, runs automatically at the end of every automatic deploy (see Verification); not a continuous monitor, only a point-in-time check.

What does not exist

  • No uptime monitoring / alerting service is configured in this repository (no PagerDuty, Cloudflare Health Checks config, or similar found in infrastructure/ or scripts/).
  • No centralized log aggregation - wrangler tail is a live stream you must be actively watching; nothing is persisted beyond Cloudflare's own dashboard retention.
  • No error-tracking SDK (e.g., Sentry) found wired into any app or Worker.

Practical implication

Detecting a production problem currently relies on: the automatic smoke test at deploy time, someone noticing (via the Cloudflare dashboard or a user report), or manually running npm run smoke:production / wrangler tail. See Incident Response for the manual detection-to-response flow this implies.