-
Notifications
You must be signed in to change notification settings - Fork 464
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Increased length of calls after upgrading Sentry #2551
Comments
Hey @DarrenStack, thanks for reporting this! We've had some performance issues (e.g. here and here) affecting our Django integration that have been fixed only recently (as in, after 1.33), as well as a lot of other bugfixes in the recent releases, so as the first step I'd recommend updating to the current version (1.38.0) and seeing if the DB times go down again. Couple questions in case upgrading doesn't help:
Sidenote regarding integration disabling, you mention that you've tried removing the Django integration -- the SDK will only actually remove it if you also provide |
Hi @sentrivana thanks so much for the quick response and all the info. The first linked ticket has a lot of overlap. On the questions:
The tip about EDIT: We have tried 1.38.0 also which has not fixed the issue |
Thanks @DarrenStack! I've been looking at the diff between 1.32.0 and 1.33.0 and nothing really stands out. One long shot is extracting some Redis connection data on each operation -- do you by any chance use Redis, e.g. to cache DB queries? Do you have any other library/framework between the app and the DB, like SQLAlchemy, or are you just using plain Django ORM? Could you add
|
Hey @sentrivana, thanks again for all the support. We are not using Redis and it's just the plain Django ORM. We have configured our cache to the database as shown here. This was initially setup for our django rate-limiting middleware. Attached are the setup logs.
|
Ok, to it is most likely the Try to set all of them to |
Thanks again for the all the ideas, they are very appreciated. Only got round to testing this now. Unfortunately I still see the same performance issues with the options all set to I'll set aside sometime in the next week to debug some more. |
Hey @sentrivana and @antonpirker, I think I found something in regards to this! We are using psycopg2-binary to support our interactions with the postgres DB. It looks to define get_dsn_parameters as a built-in method as opposed to a function so the check here fails even though it is callable. We then fall back to I realise this code was added to stop calls of |
Hey @DarrenStack, thanks for investigating and for coming back with an update! Will look into fixing this this week. |
Great! thanks @sentrivana! |
@DarrenStack The fix will be out shortly in 1.39.1. Thanks again for helping us debug! |
Yup, tested 1.39.1 and it works! Thanks for the really quick work! |
How do you use Sentry?
Self-hosted/on-premise
Version
1.33.0
Steps to Reproduce
We recently updated out Sentry SDK of our Django based product to 1.33.0 and shortly after we saw that our DB calls seem to be taking noticeably longer constantly. Everything still works fine so there's no obvious error. The same application with downgraded Sentry runs with faster calls consistently.
We are running Django 3.2.23 and the Django integration in sentry but even removing that integration has not made a difference in query time length. I have also disabled tracing with
enable_tracing
set toFalse
which did not change anything.I've looked at the changelog for 1.33.0 but can't see anything obvious that would cause this. Our application and DB are running in the same cloud services provider. Any ideas? Have only seen it for DB calls so far but not ruling out that it could be for any operation. Thanks in advance!
Expected Result
Upgraded Sentry would not effect the length of calls in the application.
Actual Result
Attached are sample performance metrics for a test set of calls we made that shows the difference between versions.
The text was updated successfully, but these errors were encountered: