Troubleshooting Crowd Performance

1. The Crowd application is slow!

1. Ensure you are running the latest version of Crowd.
2. Under Admin > Logging & Profiling in Crowd:
   • Change the com.atlassian.crowd package to DEBUG.
   • Enable profiling.
3. Replicate the performance issues you are seeing in Crowd (e.g. log out and log in, browse users, etc.)
4. Attach the resulting CROWD_DIRECTORY/atlassian-crowd.log file to your support ticket.
5. List the directories and applications active in your Crowd instance.
6. Provide rough estimates of the number of users and groups that are available in each LDAP directory configuration.
7. Provide information about the network location of any LDAP servers in respect to the Crowd server (e.g. same subnet, different networks, different states).
8. If using Active Directory, is SSL enabled?

2. JIRA/Confluence is slow!

1. Confirm that data caching is enabled in Crowd.
2. Confirm that the only crowd-integration-client JAR in the JIRA/Confluence WEB-INF/lib directory matches the version of Crowd you are running (e.g. crowd-integration-client-1.5.jar).
3. Confirm that the crowd-ehcache.xml file located in the JIRA/ConfluenceWEB-INF/classes directory matches the one in the CROWD/client/conf directory.
4. If your Crowd installation contains more than 50,000 users, review the guide at Configuring Caching for an Application.

a. JIRA/Confluence still slow?

1. Stop JIRA/Confluence.
2. Temporarily replace the WEB-INF/lib/crowd-integration-client-1.x.JAR file with the appropriate version from this issue.
3. Restart JIRA/Confluence.
4. Under Admin > Logging & Profiling in JIRA/Confluence:
   • Change the com.atlassian package to DEBUG.
   • Enable profiling.
5. Perform actions in JIRA/Confluence that are slow to respond (e.g. log out and log in, browse users, etc).
6. Attach the resulting JIRA/Confluence logs/catalina.out or stdout.log. If Confluence, also attach the atlassian-confluence.log file in the Confluence home directory (specified in the confluence-init.properties file at setup).
7. List the directories and applications active in your Crowd instance for the JIRA/Confluence application.
8. Provide rough estimates of the number of users and groups that are available in each LDAP directory configuration for the JIRA/Confluence application.
9. Provide information about the network location of any LDAP servers in respect to the Crowd server (e.g. same subnet, different networks, different states).

b. Using Active Directory?

1. Is SSL enabled?
2. Are you using nested groups (is the Use Nested Groups box checked in Crowd)?
3. If login is slow, please connect to your AD server using Apache Directory Studio and highlight the username used for this login. Provide a screenshot of this user — especially the list of memberOf attributes for this account (should contain full DNs).
4. Please also confirm that all domain controllers referenced in these groups are resolvable/reachable from the Crowd server using ping:

ping ad1.mycompany.com
ping ad2.mycompany.au

RELATED TOPICS

• Overview of Caching
• Configuring Caching for an Application
• Authorization Caching
• Configuring Caching for an LDAP Directory