Single Crowd directory for both Jira and Confluence.

[Jakob KN]

Hello folks! 

We're looking into adding Crowd as a single point of authentication and SSO for Jira and Confluence. And, to be able to manage group memberships for both Jira and Confluence directly in Crowd going forwards.

I'm considering adding a single directory in Crowd, called "Atlassian-users" which will authenticate for both Jira and Confluence. 

We've got existing users and groups in both Confluence and Jira, the majority of users and groups are identical between the two, but there are a number of users and groups which only exist in Jira/Confluence. 
Eventually, the groups will be streamlined to be the same on systems, but that's a rather large undertaking and adding Crowd takes priority over that.

In relation to this, i have a few questions:

1) If i import Users and groups first from Confluence and then from Jira, I assume that duplicate users/groups will not exist twice in the "Atlassian-users" directory? 

2) Groups and users that only exist in Jira will only be created in the "Atlassian-users" directory, but not directly in Confluence?
(and vice versa for groups/users only existing Confluence)

3) Is this a viable approach, or is it better to have a two separate directories in Crowd, for Jira and Confluence? 

We're on Data Center for both Jira, Confluence and Crowd.

Any answers, tips or advice is much appreciated.
Thanks in advance!

[Arnav Ghatage · Accepted solution]

Hello Jakob!

I'd be happy to help with your questions regarding using Crowd as a single point of authentication and SSO for Jira and Confluence.

1) If i import Users and groups first from Confluence and then from Jira, I assume that duplicate users/groups will not exist twice in the "Atlassian-users" directory?

>> If you import users and groups from Confluence and then from Jira into the "Atlassian-users" directory in Crowd, any duplicate users or groups should not exist twice in the directory. Crowd should recognize and merge the duplicate entries during the import process.

2) Groups and users that only exist in Jira will only be created in the "Atlassian-users" directory, but not directly in Confluence?
(and vice versa for groups/users only existing Confluence)

That's correct, but as you've synced the Crowd directory to confluence and jira, all the users synced in the Crowd directory will be visible in Jira and Confluence.

3) Is this a viable approach, or is it better to have a two separate directories in Crowd, for Jira and Confluence?

Using a single directory in Crowd for both Jira and Confluence is a viable approach and is actually recommended by Atlassian. It simplifies user management and allows for centralized control over authentication and authorization across the two systems. However, it's important to ensure that the user and group data is kept in sync between Jira and Confluence to avoid any issues.

I hope this helps! Let me know if you have any further questions or concerns.

P.S - I work for miniOrange, the exclusive Crowd SSO vendor on the Atlassian Marketplace

Thanks,

Arnav Ghatage