Skip to content

Capacity or rate limits

Check what the provider actually reported before retrying manually — and let AIgency's automatic recheck do the work if you're in a run.

A capacity or rate-limit error from a provider is a temporary, external condition — not a bug in AgentDock, and not necessarily something that needs manual intervention.

In a normal session

Check Provider usage for whatever signal the provider CLI actually reports — an exact quota value where parseable, or a general capacity/cooldown signal otherwise. Retry once that clears rather than immediately hammering the same request.

In AIgency

AgentDock Pro

Capacity recovery is part of AIgency.

A task blocked this way moves to waiting_capacity and is rechecked automatically — on a known reset time if the provider reported one, or on a bounded backoff (roughly 2, 5, 10, then 15 minutes) if not. This never counts against the task's retry budget, so there's usually nothing to do manually — see Capacity recovery. Forcing a manual retry before the recheck fires just burns another attempt against the same still-exhausted limit.