Description
We have a consumption plan function app that has started producing spikes of this exception in both our production and UAT environments. I believe this is related to #10580 and/or #11119 . The function appears to be well behaved but then the language worker is killed. I've seen this proceeded by a trace like either "Language Worker Process exited. Pid=144." or "python exited with code 137" (indicative of a sigkill). For some reason the runtime(?) appears to continue pick up triggers and create invocations which are then almost immediately failed with the above exception. During this time the memory usage of the function worker usually spikes. This can continue for several minutes and chew through a lot of messages, causing some to be moved to poison.
Questions:
- Why is the language worker being killed?
- Is this the same underlying bug as Unexpected invocation timeout notifications after terminating the .NET Isolated worker process #10580 or something else?
- What can I do to prevent this?
Investigative information
Please provide the following:
- Timestamp: 6/23/2025, 4:42:16.8650252 AM
- Function App version: 4
- Function App name:
- Function name(s) (as appropriate):
- Invocation ID: 74e03d42-321e-49e4-92cf-1d6b179afc23
- Region: West Europe
Related information
Provide any related information
- Programming language used = Python 3.12