err:frontline:capacity:deployment_offline
The gateway resolved this hostname to a deployment that is intentionally offline. It was stopped, superseded, or its project was cancelled, so there is no running compute to serve the request. The gateway returns 503 Service Unavailable.
This is not the transient gap you see mid-deploy or mid-scale (that is no_running_instances). An offline deployment does not come back on its own.
What to do
- If the project is still active, deploy again or promote a running deployment so the hostname points at live compute.
- If the project was cancelled, re-subscribe to Compute to bring it back.
- If you hit your spend limit, you will see
spend_limit_reached instead, which resumes automatically once you raise the budget.