Parallel code context ops crash container #276
Description
Encountered while working on #266.
The code-interpreter-workflow.test.ts was permanently timing out for tests creating and deleting multiple contexts at once. Since I was doing that in parallel, I tried changing it to send the requests one at a time instead and that seemed to fix the issue. I guess it could be a bug in the container image on how it handles that.
Only logs I was able to get:
Good next steps to test is to go to those tests and change to be parallel again, increase how many contexts they create and deploy to a test worker with debug logs and see where it might be happening.