Java min and max for 3000 issues

[Kevin L]

Hi

we have had two system lockups where we noticed that we could not get into the portal. Would just hang.  Able to putty into server.  Received alerts that there was a high load on server.  

We have about 3000 issues in the system after about 5 months of use.  Atlassian support agent said memory of 1000/3000 (min/max) is too high.  Other agent said too low to increase to 4gb.  Not sure which is right.  

Sure coudl use help.  We will be enabling GC logs tomorrow for more insight.  Also will make sure all plugins are up to date. JIRA core is 7.3.3

[Nic Brough -Adaptavist-]

It does sound like a GC problem, so enabling logs would be good.

3,000 issues doesn't need 3Gb RAM for the most part, so having it too high might be leading to long garbage collections.  Increasing RAM if it is a GC error might just reduce frequency but increase their length.  You'll want to get to the bottom of what the load and memory usage is.

[Kevin L]

The guidance in their article says to use 768mb max for under 5k issues. But when I look at memory, I have seen used memory over 2gb.  right now 10pm ET which is non peak time I see:

54% Free (Used: 1417 MB Total: 3111 MB) 

How do I know what to reduce it too min and max?

[Nic Brough -Adaptavist-]

 I would try reducing both to 768Kb as the article suggests, but do it in steps of 256Kb or 512Kb and monitor it carefully to see how the pattern of problems changes.

However, changing the memory settings might not be the right thing to do.  It might improve it in terms of impact, and it could make it worse, but I doubt it will make it go away completely. 

It feels to me like you have a memory leak and/or a performance issue, and you need to find out what that is.  Fixing that will stop GC from freezing your system