Massive User Directory Performance Issues

Brandon Viertel January 13, 2020

TL;DR: Massive user base from LDAP, massive performance issues. 

We have a massive LDAP connected user directory that we use for our instance, with about 260,000 users. Because the user base is so large, we are experiencing performance issues with any action related to it. Items such as the assignable user picker encounters a timeout when using it, and we've clocked it at 1.7 minutes using the REST API. Whenever this is used, it eats a thread for that amount of time, and if 10 users select that at the same time, it pretty much crashes our instance eating up all the CPU.

A recent issue we had encountered and needed to restart our Jira instance, was that a thread was blocked on the following code: 

com.atlassian.jira.crowd.embedded.ofbi.IndexedUserDao.update

This once again relates to crowd, which is the user management. We don't use a crowd directory here, as we only have the internal and LDAP connected directories. 

This issue is becoming more and more critical, and we are wondering how others are managing a larger user base, and if anyone has any suggestions to the issues we are encountering. 

We have already ensured that there is enough hardware behind it, however it appears it will consume ANY amount sent it's way. 

We are running Software and Service Desk, all on Jira Server.

THANK YOU!

1 answer

0 votes
Gonchik Tsymzhitov
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
January 3, 2021

Hi! 

Please, review integration with LDAP and set the filters and check latency

https://confluence.atlassian.com/kb/how-to-write-ldap-search-filters-792496933.html

Suggest an answer

Log in or Sign up to answer