Delayed Transaction Counting

Dear Sentry and fellow On-Prem members,

I seem to have encountered a weird issue recently in that all of the transaction counting metrics on my sentry instance are all delayed by several hours:

As you can see I have had a couple of Errors recently and these have been logged however what can also be seen is that there are no transactions logged. Strangely the last metric which was several hours before the current time will continue to count up until the next hour ticks over. It is as if it is on the wrong time zone however the server time zone, user/client time zone and local machine time zones are all setup correctly.

Details to note:
I have the PostgreSQL instance running on another machine due to the high amount of disk space required (This is working correctly as expected). I am also using the GitHub integration however other than that there is nothing custom or otherwise modified.

If anyone could shed some light on where to look to solve the issue that would be greatly appreciated. Thank you.

PS. Sentry Team, I am loving the platform so far, Keep up the awesome work!

1 Like

I’d suspect an overloaded Kafka instance first. If that’s not the case you may wanna check this: https://develop.sentry.dev/self-hosted/troubleshooting/#workers

PS. Sentry Team, I am loving the platform so far, Keep up the awesome work!

Thanks a lot for the kind words! :heart:

Hello BYK,

So, I have tried resetting/aligning kafka however that does not seem to have done it. looking at kafka and worker logs neither seem to be throwing any obvious issues. Equally I am not seeing anything to suggest the workers are struggling.

Do you have any other suggestions by any chance?
If it helps I am of course happy to provide any logs just let me know which ones you would require.

Thank you.

Okay so I tried the recovery method on the link however this did not resolve the issue.
I have now resorted to using the nuclear option here and that seems to have done the trick.
Thank you for your assistance.