Hello everyone! We’ve been experiencing some issues with the Metaflow production UI becoming increasingly slow. It seems to be related to some slow sql queries to our expanding metadata table, which is now around 6GBs . We have tried increasing the resources around this database, but I don’t think this is a feasible long-term solution.. was curious if other users face similar issues and what your team suggests? Thanks in advance for your insights! versions: Application version: v1.1.2 Service version: 2.4.2--

There have been some significant query optimizations on the UI service version since 2.4.2 at least, so if you are able to update that to a more recent version it might alleviate some of the issues. The table size isn't unreasonable though so I'm wondering if there is possibly something else going on as well. Are you using one of the suggested deployment templates that either uses the migration service, or

run_goose.py

to execute possible migrations? This would be on the metadata service side. The UI performance is very index dependent so if some are missing, it might explain the slowdowns. Lastly the UI has also had some bugs fixed with api access patterns since 1.1.2 so that might also be worth updating.

@astonishing-jewelry-18059 this looks very similar to an issue I experienced. Try restarting the ui-backend container.