Skip to content

Use asyncio.to_thread to avoid DeadlockError while sending exception to Sentry #267

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Natim wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Use asyncio.to_thread to avoid DeadlockError while sending exception to Sentry #267

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions sentry/interceptor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,7 @@
)

with workflow.unsafe.imports_passed_through():
import asyncio
import sentry_sdk


Expand Down Expand Up @@ -43,7 +44,7 @@ async def execute_activity(self, input: ExecuteActivityInput) -> Any:
if is_dataclass(arg) and not isinstance(arg, type):
scope.set_context("temporal.activity.input", asdict(arg))
scope.set_context("temporal.activity.info", activity.info().__dict__)
scope.capture_exception()
await asyncio.to_thread(scope.capture_exception, e)
raise e


Expand Down Expand Up @@ -71,7 +72,7 @@ async def execute_workflow(self, input: ExecuteWorkflowInput) -> Any:
scope.set_context("temporal.workflow.info", workflow.info().__dict__)
if not workflow.unsafe.is_replaying():
with workflow.unsafe.sandbox_unrestricted():
scope.capture_exception()
await asyncio.to_thread(scope.capture_exception, e)
Copy link
Member

@cretz cretz Dec 19, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is not legal code from a workflow, and as you saw, is not supported by our event loop (by intention). Two things here...

First, Sentry is on the verge of a 3.x SDK which will include async support (per getsentry/sentry-python#2824 (comment)). They appear to have released a few alphas already.

Second, not sure we should assume they want to use the default executor and you're going to have to use an intentionally-undocumented feature called "unsafe extern" here (we use this for a similar purpose for OTel at https://github.com/temporalio/sdk-python/blob/main/temporalio/contrib/opentelemetry.py). I would suggest doing the following:

  • Have the constructor of SentryInterceptor accept an optional ThreadPoolExecutor, and capture the current event loop as a private attribute on the class
  • Have a def _capture_exception(self, scope: sentry_sdk.Scope, exception: Exception) that calls the loop's run_in_executor with the known thread pool executor for scope's capture_exception
  • Pass the SentryInterceptor to _SentryActivityInboundInterceptor's interceptor so it can use this
  • On workflow_interceptor_class, call input.unsafe_extern_functions.update("__temporal_sentry_capture_exception", self._capture_exception)
  • In _SentryWorkflowInterceptor instead of what's called today, call temporalio.workflow.extern_functions()["__temporal_sentry_capture_exception"](scope, e)

But would understand if there isn't value in doing all of this if it's just going to be outdated shortly with v3 SDK

raise e


Expand Down