Issue Summary

On a very large instance (e.g. 10,000 projects and 2000 roles), JQL search by text can be extremely slow in some settings when the cache is cold (e.g. server restarted).

It can take > 20 mins.

This is reproducible on Data Center: yes

Steps to Reproduce

1. Provision Jira 9.x with a very large dataset (e.g. 10,000 projects and 2000 roles)
2. Login as admin and JQL search text ~ "testing" or any other text on two separate browser windows
3. JQL does not load ( > 20 mins)

Expected Results

JQL returns in a few seconds, or less than a minute at max (first-time access).

Actual Results

JQL does not return, even after 20 mins in. This issue is related to JRASERVER-71695.

Attached is the stack trace.

Workaround

Limit the number of actors assigned to a given project role by:

1. Archiving unused projects, if any
2. Splitting roles - e.g. distributing users from large "Developers" roles into "Developers", "Testers", "IT Ops" etc.
3. Limiting the number of projects a role can view