-
Notifications
You must be signed in to change notification settings - Fork 41
Description
Describe the bug
Uncaught exceptions in our deployed flask application aren't sent to the BugSnag platform on version v4.8.0 of the library. The most recent release before this, v4.7.1, doesn't have an issue. We've currently pinned our app to use 4.7.1 to get around this.
Steps to reproduce
Unsure. I cannot reproduce locally when swapping between versions -- when run locally and swapping between 4.8.0 and 4.7.1, both versions propagate uncaught exceptions to the the BS platform.
Our Flask application is deployed in a lambda environment through the serverless framework. I'm assuming something about how we're configured is the issue. I'm hoping a maintainer more familiar with the library can reason through the delta between 4.7.1 to 4.8.0 and point out what might be different that would allow a behavioral change with the BS SDK when none of our application code changed.
Environment
- Bugsnag version: 4.7.1 and 4.8.0
- Python version: 3.10
- Integration framework version:
- Flask: 2.3.3
We configure our BS client as one of the last lines of our our flask app __init__ like so:
bugsnag.configure(
api_key=self.config.get("BUGSNAG_API_KEY", None),
app_version=self.config.get("GIT_SHA"),
app_type="internal_api",
release_stage=os.getenv("AWS_ENVIRONMENT"),
send_environment=True,
)I took the diff of changes between the 4.7.1 and 4.8.0 release (link here) and I think the issue could be coming from one of these two commits (one, two), but otherwise unsure.
It feels like it could be the endpoint isn't being configured correctly, as that seems like the new add between the two versions, but I don't know how to rule that out locally (or where to put breakpoints to check).