Architecture

Status checks are performed server-side by a Supabase edge function (status-monitor) running in the same region as the RB infrastructure — not from your browser. This eliminates CORS false positives that plague client-side status pages.

The monitor runs on a pg_cron schedule every 15 minutes. Results are stored in rb_status_checks in Supabase. This page fetches the latest results from that function on load and every 5 minutes.

What Each Check Actually Does

Netlify sites — HTTP HEAD request. Green = 2xx/3xx response. Red = 5xx or timeout.
Supabase — Authenticated REST API ping. Verifies database is accepting connections.
Resend / Shopify / Anthropic — Parses their official Atlassian status JSON API (/api/v2/status.json). Reads the indicator field directly — not the webpage.
Agents — Queries cron.job_run_details for scheduled agents, and performs a HEAD request to the edge function URL for Make.com-triggered agents.

When a Break Is Detected

A single failed check does not trigger an alert — this prevents false alarms from transient blips.

Alert threshold: 2 consecutive failures. If a service fails two checks in a row (30 minutes apart), the monitor sends an alert email immediately to saul@rachelbloom.com and rachel@rachelbloom.com naming exactly what failed and why.

A recovery email is sent automatically when the service comes back online on the next check cycle.

Status Indicators

Operational — responding normally
Degraded — responding but with issues (e.g. Shopify reports a minor incident)
Down — not responding or reporting a major incident

The hero bar at the top reflects the aggregate state. If 3 or more services are down simultaneously it turns red and reads "Service Disruption Detected."