feat(request): add timeout and retry logic #44
Conversation
indexclient/client.py
Outdated
|
|
||
| def indexd_request_wrapper(func): | ||
| def retry_logic_with_timeout(*args, **kwargs): | ||
| kwargs.update({"timeout": 60}) |
There was a problem hiding this comment.
did you mean to use set_default()? The update here overwrites any given timeout.
There was a problem hiding this comment.
Good catch. Thanks!
| return resp | ||
|
|
||
| @indexd_request_wrapper | ||
| def _post(self, *path, **kwargs): |
There was a problem hiding this comment.
I'm still having concerns adding retries on POST, PUT or DELETE - it is totally fine to add timeouts, but retrying means any future IndexD APIs must be strictly repeatable (calling with the same args for more than one times ends up with identical results), if current APIs are already repeatable. What was the initial motivation to retry on e.g. POST please?
There was a problem hiding this comment.
I have a ticket to copy and index million files. We have to make sure that all the new files need to be indexed in indexd. For safety, I think I will only add timeout for POS, PUT and DELETE and change the logic in my code. Thank you for your comment!
Fixed in 5f39c8b
f108389 to
5f39c8b
Compare
Description about what this pull request does.
Add timeout and retry logic for all requests made to indexd api
Please make sure to follow the DEV guidelines before asking for review.
New Features
Add timeout and retry logic for all requests made to indexd api
Breaking Changes
Bug Fixes
Improvements
Dependency updates
Deployment changes