Making Jira more resilient towards heavy users

[bschmi]

Hi Community,

from time to time we see users in our setup which tend to overload Jira using their browser.

We are on Jira 8.5 on a single server (no data center) and use a reverse proxy.
RAM 64 GB, Heap 16 GB, 4 CPU Cores

Users can easily push the load on CPU cores by hammering on reload button. We also see some heavy REST API users.

Does somebody have restrictions, on reverse proxy basis or using iptables (or similar) in place preventing such overload and could share some insights/values?

If not, the next for us would be to google around suitable tools (probably, iptables or a Apache module) to implement rules preventing one single user can block the whole Jira instance.

I am happy about any hints and tricks you could share.

Thanks,
Birgit

[Nic Brough -Adaptavist-]

First, I would talk to your users, they need to understand that "hammering on reload button" is what is making it slow and they need to learn to stop doing it.

There are a couple of things to look at.  I would be considering moving to data centre if you have severe loading issues.  There are two things you could do with DC that might help:

Rate limiting is built into DC - https://confluence.atlassian.com/adminjiraserver/improving-instance-stability-with-rate-limiting-983794911.html - you'll be doing all the config in a system that understands what parts of Jira you would want to limit

With a DC, you have options like being able to dedicate nodes to REST, so REST activity won't affect your UI users.

If DC is not an option, then yes, I'm afraid you'll be looking at building your own solutions.  I've done something simple with Apache's mod_ratelimit in the past, a very simple "stop developer (who had already been given a better option, but was too lazy to change) hammering REST".  I suspect the proxy is the best place to do this.