Status page

A public, unauthenticated page at /status that mirrors the platform's deep healthcheck. Linkable from monitoring tools and from enterprise security questionnaires.

What's reported

The page reads the live /api/v1/health endpoint and renders:

• Overall summary: All systems operational, Degraded: non-critical service issue, or Down: core service unavailable.
• Component probes: database, rate-limit store, models server. Each shows its current state and last-probe latency.
• Circuit breakers: open / closed / half-open state for the retrosynthesis circuit (Paperspace) and the PubChem vendor-lookup circuit. An open breaker means the upstream is unhealthy and the platform is degrading gracefully rather than 503-ing.

The page refreshes per-request, so a hard reload always reflects current state. No client-side caching.

Wiring monitors

The same /api/v1/health endpoint that powers the page is suitable for external uptime monitors:

• Keyword "summary":"ok": fully healthy.
• Keyword "summary":"degraded": degraded but functional.
• HTTP 503: fully down (core dependency unavailable).

We use this internally with UptimeRobot and recommend the same setup for enterprise customers wiring their own monitoring.