System check shows... "Oops… there was a problem during the request"


Just after a small bit of help or advice… We have a Matomo installation that is having a few issues at the moment. Any help or pointers would be much appreciated (I’m fairly new to Matomo). Currently the error in the admin console (where the system-check panel should be) is:

Oops… there was a problem during the request. Maybe the server had a temporary issue, or maybe you requested a report with too much data. Please try again. If this error occurs repeatedly please contact your Matomo administrator for assistance.

Selecting the “System Check” option in the Diagnostics menu causes the Matomo and nginx containers to exit and restart (this runs in AWS ECS with an AWS Aurora MySQL database).

The environment also has some issues where archiving is very slow - possibly linked.

The Matomo version is: 4.0.5

Currently the archiving container / process seems to be hung. The last message in the archive log is:
Found archive with intersecting period with others in concurrent batch, skipping until next batch
This message is repeated every 90 mins or so.

Apologies for the long post - but hopefully this gives some clues. Let me know if there’s any additional info I can provide or any useful diagnostics I can run (DB queries, etc.).

Thanks in advance.

Hi Gareth,
could you maybe check this FAQ

in particular:

  • If the error persists, or you’re getting a white screen/page, it indicates that there is a problem on the server: check in your web server error log file for an error message.
  • When you find an error message in your server error log, then search for this message in the Matomo FAQ.

maybe you find an error message there?

Hi @matthieu

Thanks for the update. Yeah, I think this was all related to our UsersFlow issue that I’ve been chatting with @diosmosis about. It was all happening around the same time - and since we’ve temporarily deactivated the UsersFlow plugin, all has been well with archving and we no longer see this error either.

The messages in the logs at the time were “MySQL server has gone away” - though I think more likely to have been overloaded with UsersFlow archiving records, causing System Check queries to time out rather than the ‘server’ being “away” as such (the setup is actually ‘serverless’ in AWS).

Also where I mentioned the system being ‘locked up’ with Found archive with intersecting period... messages - I suspect this was the archiver trying to run with 10’s of millions of records in the matomo_archive_invalidations table - and queries running slow (slower than records were being cleared).

We’re making plans to reintroduce the plugin so will test the System Check page when we do that. Also, we’ll monitor the number of rows in the matomo_archive_invalidations table and check that hourly archiving clears this table after each run. :crossed_fingers:

Thanks again.