Issue Summary

Syncs between Confluence & Crowd (with delegated LDAP directory) can cause longer than expected sync times due to a large number of groups. This is due to difference in the DAO models between Crowd and Confluence and the volume of groups as required by the sync.

Steps to Reproduce

1. Integrate Confluence and Crowd
2. Set up a Confluence application in Crowd
3. Create a directory and add a delegated connection to AD with ~ 40k groups. Associate this directory with the Confluence application in Crowd
4. Sync the directory from Confluence

Expected Results

1. Sync times will remain relatively low and will not take longer over time.

Actual Results

1. Sync times gradually increase over time as groups are added to the data set.

Workaround

Confluence doesn't come with indexes on the lower_group_name, but adding one can certainly improve processing speed:

For Postgres:

CREATE INDEX idx_lower_group_name ON public.cwd_group USING btree (lower_group_name);