Highload optimization

Sentry performance test shows me that maximum throughput is ~100 messages per second. I used standard docker installation. After testing Sentry continues to process queues. It continues too much time. Could anybody tell best practices to improve Sentry performance?

I run sentry on kubernetes. I used redis broker. You can launch more worker containers in your pod. Increase web processes to 5 in your sentry-web container.

I hope sentry replaces celery-beat with redbeat. You can do it too. It is a better scheduler and stores state in redis. From the looks of it, sentry receives the crash report and queues it for processing.

Now based on your web requests frequency, scale the pods. Make sure your filestore and database (postgres recommended on RDS) are highly scalable.