Recently we started to test YT and it's performance, but later on we'd like to move the organization's data on it. However we recently noticed, that after the full migration of our Confluence spaces' data the knowledge base's performance has dramatically dropped. (This is roughly 200 projects and ~6.5k articles, about 10GB of data.)
- To list the knowledge base tree takes roughly 10-12 sec
- To display a random article takes 5-7 sec
According to our analysis the API's answer of the server takes roughly 7 sec with a 2 sec period of idling, which would be not good but in an acceptable range. But at the same time from this point on the client-side the building of the articles to display takes extra 5-7 sec on an i5 with a 100-150% browser CPU load.
The question is, could this be optimized somehow now, like limiting the size of pagination or the displaying elements, or do we have to wait for a future patch/update on the 2020.6 version?
Facts about the server and of our infrastructure:
- 12 vcpus
- 8GB memory
- The sole Docker instance that is running on the host is YT
- The heap size is 6GB
- The MaxMetaspaceSize is 1GB
- The DB size is 10GB with just the articles in it
- YT running via HTTPS behind a HAproxy
Thanks for the answer in advance.