Fix the telemetry exception handling for daemon mode #60
Conversation
In daemon mode, the request is made in subprocess effectively causing a huge set of stacktrace in terminal as provided in iterative/studio#11282. This handles the exception properly along with timeout for the telemetry call. Closes iterative/studio#11282
src/iterative_telemetry/__init__.py
Outdated
| '{self.url}', | ||
| params={{'token':'{self.token}'}}, | ||
| json={payload}, | ||
| timeout=2 |
There was a problem hiding this comment.
I think it should be safe to make it 10+ seconds?
There was a problem hiding this comment.
I used the default from the timeout we already had for thread based or normal method of sending the request. I am not sure if we want to have timeout upto 10 seconds delay to close a program when program already failed for the user.
There was a problem hiding this comment.
if we want to have timeout upto 10 seconds delay to close a program when program already failed for the user.
are we waiting for the process to exit / complete on the user program exit?
There was a problem hiding this comment.
Looks like we have close_fds=True, so, we won't have the blocking exit. Will update the timeout to 10 seconds.
There was a problem hiding this comment.
Well, close_fds=True does not close stdin, stdout and stderr streams. So it might end up blocking. But that's a separate issue.
In daemon mode, the request is made in subprocess effectively causing a
huge set of stacktrace in terminal as provided in
https://github.com/iterative/studio/issues/11282.
This handles the exception properly along with timeout for the telemetry
call.
Closes https://github.com/iterative/studio/issues/11282