How do I tune Sentry Selfhosted when it seems slow

We have a Sentry hosted on a VM with 6 Cores and 64 GB RAM. This is the first time we have deployed it and do not have the expertise to maintain it. There is some information available in the forums and online but it is hard to put it all together and make sense of it. In our setup we are seeing a few problem.
Sentry web is slow and the web interface takes time to load.
If I stop Sentry and start it I do not get the web interface and continue to see "502 bad Gateway”(NGINX). It usually requires a few restarts before I get access to the web interface.
The search queries can take a long time to complete. Some of the queries fail all together with an error on the web interface. The logs are only helpful if you understand what to look for to indicate a bottleneck.
I would like to tune the setup to meet our requirements just to give you an idea. We have 9.6 M transactions to date and we generate depending on time of day between 10000-150000 transactions an hour. At the moment our Docker volumes are showing usage of 421 GB. Most of it is being used by Postgres (365 GB) and Kafka (34GB).

Sounds like a great case for using our hosted offering. :slight_smile:

1 Like

I second @chadwhitacre but in case you want to try harder, you may find some pointers at https://develop.sentry.dev/self-hosted/troubleshooting/