Skip to main content

​
Timeout Policies

Timeout policies protect workflows from running indefinitely. A timeout defines the maximum amount of time a workflow or operation is allowed to execute before it is considered to have exceeded its execution limit. Timeouts improve system stability by preventing stalled or excessively long-running tasks from consuming resources indefinitely.

​
What is a Timeout?

A timeout specifies the maximum execution duration. Conceptually:
If execution exceeds the configured limit, the timeout policy is applied.

​
Why Use Timeouts?

Not every operation finishes successfully. Sometimes tasks become stuck because of:
  • unresponsive services
  • infinite loops
  • blocked network requests
  • external API delays
  • infrastructure failures
  • unexpected system behavior
Timeouts prevent these situations from affecting the entire workflow.

​
Execution Timeline

A timeout measures elapsed execution time.
If the task finishes before the limit, execution continues normally.

​
Timeout Policy

BindAI provides a timeout policy object. Example:
The timeout specifies:
  • maximum execution time
  • behavior after timeout

​
Successful Completion

If execution completes within the allowed duration:
The timeout policy has no effect.

​
Timeout Triggered

If execution exceeds the configured limit:
The workflow responds according to the configured policy.

​
Failing the Workflow

A common configuration is to fail the workflow immediately.
This prevents additional processing after an unrecoverable delay.

​
Continuing After Timeout

Some workflows may choose not to fail immediately. Conceptually:
This behavior depends on workflow requirements and timeout configuration.

​
Timeout vs Retry

Timeouts and retries address different concerns. Both policies are often used together.

​
Timeout vs Loop

Loops intentionally repeat execution. Timeouts prevent excessive execution time.
A timeout provides protection if loop conditions never become false.

​
Common Use Cases

Timeout policies are useful for:
  • external API calls
  • AI model inference
  • document processing
  • long-running workflows
  • scheduled jobs
  • network communication
  • background automation
Any operation with uncertain execution time can benefit from a timeout.

​
Combining Policies

Production workflows often combine multiple reliability mechanisms. Example:
This allows temporary problems to recover while preventing indefinite execution.

​
Monitoring

Timeout events should be monitored in production systems. Useful information includes:
  • execution duration
  • node identifier
  • workflow identifier
  • timeout threshold
  • retry attempts
  • error details
Monitoring helps identify performance bottlenecks.

​
Best Practices

  • Set realistic timeout values.
  • Use shorter timeouts for external services.
  • Combine timeouts with retry policies.
  • Monitor timeout frequency.
  • Avoid excessively large timeout limits.
  • Design workflows so interrupted tasks can be restarted safely when appropriate.

​
Summary

Timeout policies protect BindAI workflows from stalled or excessively long-running execution. By limiting execution duration, they improve system stability, prevent resource exhaustion, and provide predictable behavior when operations fail to complete within acceptable time limits.