AI Agents / Agent orchestration and reliability
AI Agents: Fallback policy
Define when to retry, switch models or return partial results based on failure class.
Environment and inputs: Application-managed queues and worker orchestration; backend scheduling and recovery guarantees vary.
When to use it
Additional workers create coordination and resource costs as well as parallelism. Start with explicit artifact ownership and a shared completion definition, then test saturation and partial failure using controlled mock dependencies.
Check the result
Trigger a capability failure and verify fallback does not silently remove a required tool guarantee.
Save and compare results
Save the test input, configuration and result. Fix the failing boundary, repeat the same test and compare it with the saved working configuration.
Scope and limits
These exercises have not been tested on your systems. Check feature support and configuration names against the installed release and linked documentation.