Status
Is Margin up?
No surface is reporting a fault; 1 could not be checked from here. This page covers the three surfaces a pilot depends on. It does not watch your own application, your model provider, or the individual runs in your console.
Checked 2026-10-20 05:41 UTC. State comes from a direct check taken when this page loaded.
Website
Operational
trymargin.io and the pages you read before signing up.
- Last 24h
- not measured
- Last 7 days
- not measured
API and database
Operational
Sign-in, the console, and every read of your measured spend.
- Last 24h
- not measured
- Last 7 days
- not measured
Run worker
Not checked
Where your agents actually run: measurement and face-off jobs.
- Last 24h
- not measured
- Last 7 days
- not measured
Live monitor history is not connected on this deployment, so these are direct checks taken now — not one-day or seven-day uptime. The run worker runs on a separate host and cannot be checked from here.
What this page watches — and what it does not
The website, the API and database behind sign-in and every measured figure, and the worker that runs your agents. Those are the parts a failed run can be blamed on. A slow or failing model provider is not on this page: your own console shows which calls failed and what they cost, and provider status is published by the provider.
If something breaks
Write to subh@trymargin.io. It reaches the founder, not a queue, and we acknowledge it within one business day. If a run fails because of us, we re-run it at our cost and it does not appear on your invoice.
What happens in the first hour is written down rather than improvised: who reads the mailbox, where alerts land, and how a failed customer run is caught before you report it. If you want that page before you sign, ask and we will send it.