Status
Checks run from outside our network by Gatus. The embedded page below is the live view; open status.ai.ml for history, incidents and the RSS feed.
What is checked
- Gateway api.ai.ml/healthz every 10 s from three regions
- Control plane api.ai.ml/v1/models (catalog snapshot served)
- Console and docs login page and docs home
- Provider routes a real completion through each provider route every minute; failures show which host is degraded and whether failover absorbed it
Per-model uptime and latency over 30 days are on each model page; incidents are posted on the status page with a postmortem within five working days.