Retries & Timeouts
Keep workflows resilient by automatically recovering from temporary failures and preventing stuck tasks.
External systems fail.
APIs slow down.
Webhooks timeout.
Dependencies become temporarily unavailable.
Retries and timeouts help workflows recover without creating hidden operational failures.
Automatic retries
Cascades can automatically retry tasks when temporary failures occur.
Examples include:
- API rate limits
- temporary outages
- network interruptions
- webhook failures
Teams can configure:
- maximum retry attempts
- retry intervals
- backoff policies
This helps prevent unnecessary workflow failures.
Backoff protection
Retries should not overwhelm external systems.
Backoff policies gradually increase retry delays to reduce pressure on:
- APIs
- databases
- external providers
- internal systems
This helps avoid repeated failure loops.
Timeout protection
Some tasks should not run indefinitely.
Timeouts help stop workflows that become stuck because of:
- slow APIs
- failed integrations
- waiting systems
- broken dependencies
This prevents stalled tasks from blocking worker capacity.
Failure escalation
When retries fail or timeouts are exceeded:
- workflows can pause
- alerts can trigger
- operators can investigate
- recovery workflows can begin
See Failure Recovery.
Operational visibility
Teams can review:
- retry attempts
- timeout failures
- execution logs
- task history
This makes workflow reliability easier to manage at scale.
Verification support
Retries and failures remain visible in execution records and proof artifacts.
Teams can review what happened during workflow execution without losing historical context.