Retries
Failed AIgency tasks are retried automatically up to a limit, with capacity-aware backoff for provider rate limits.
Individual task attempts can fail — a rate limit, a transient error, an agent giving up. AIgency retries automatically within limits you control.
AgentDock Pro
Retry handling is part of AIgency, available only in AgentDock Pro.
Retry accounting
Each task tracks a retryCount against a maxRetries limit, and AIgency keeps a history of every prior attempt for the task so you can see exactly what was tried before the current attempt.
Capacity-aware backoff
When a failure looks like a provider rate limit or capacity error rather than a genuine task failure, AIgency classifies it as such and backs off automatically, retrying once capacity is likely to be available again rather than burning through the retry budget on errors outside the task's control. This is reflected in the task's waiting_capacity status — see Execution for the full status lifecycle.
Related
- Agent assignment — fallback agents and
allowReassignOnFailure, for when retrying the same agent isn't the right response to a failure.