[Feature Request] Configurable Deadlock Detection Timeout in Python SDK

[SanilK2108]

Is your feature request related to a problem? Please describe.

Currently, the Temporal Python SDK enforces a fixed 2-second deadlock detection timeout for workflows. This behavior cannot be configured in production: the only way to disable it is to run in debug_mode or set TEMPORAL_DEBUG.

This creates friction for production workflows that legitimately need longer CPU-bound or blocking sections. Unlike the Go SDK, which allows configuring WorkerOptions.DeadlockDetectionTimeout, the Python SDK forces developers to restructure code even in cases where a slightly longer threshold would be sufficient.

Describe the solution you'd like

Introduce a configurable deadlock detection timeout for the Python SDK, similar to Go’s DeadlockDetectionTimeout option.

Example proposal:

worker = Worker(
    client,
    task_queue="my-task-queue",
    workflows=[MyWorkflow],
    activities=[...],
    deadlock_detection_timeout=timedelta(seconds=5)  # new option
)

Default remains 2 seconds (to preserve current behavior).

Users can override the timeout per worker in production.

Setting None (or 0) could fully disable the check, mirroring debug mode.

Additional context

[SanilK2108]

I’d be happy to contribute this feature and open a PR to add configurable deadlock detection timeout support to the Python SDK.

[Sushisource]

@SanilK2108 We would welcome a PR! Actually, after some discussion with the team we're pretty sure we want to keep this limited purposefully. If you're doing things that extend the timeout beyond those values, there's a nearly 100% chance that you shouldn't be doing those things in a workflow context.

IE

workflows that legitimately need longer CPU-bound or blocking sections

There is pretty much never a "legitimate" need for this. Pushing that code down into a local (or nonlocal) activity is the right thing to do.