Summary

When a plan is created that is based on a JQL query that does not contain any project ids and yields a very large number of issues it is possible to cause an out-of-memory exception.

This is because the getProjectIdsViaExhaustiveSearch function attempts to determine the project IDs by running the JQL query and then iterating over the results to find the set of projects in the plan. However, the search is performed without any limit on the number of results and this means that if there are a lot of results then memory can be exhausted during the processing.

Environment

• Large instance with > 1M issues

Steps to Reproduce

1. generate 1M issues
2. create Portfolio Plan with empty board filter
3. load this board repeatedly

Expected Results

• there should be a limit of issues can be loaded in a single board
• there should be a validation check for creating board with empty filter

Actual Results

Jira consumes large heap and eventually OOME

Notes

Suggested resolution is to have to page over all the results (with the page size being that of the issue limit) to ensure that issues are processed in manageable chunks.

Workaround

Retrieve the empty JQL plans using the following SQL query

select "p"."TITLE", "is"."SOURCE_TYPE", "is"."SOURCE_VALUE","rv"."SAVED_FILTER_ID", sr.reqcontent 
from "AO_D9132D_ISSUE_SOURCE" "is" full 
join "AO_60DB71_RAPIDVIEW" "rv" on cast("is"."SOURCE_VALUE" as int) = "rv"."ID" 
join searchrequest sr on (sr.id = "rv"."SAVED_FILTER_ID" or sr.id = cast("is"."SOURCE_VALUE" as int))
join "AO_D9132D_PLAN" "p" on "p"."ID" = "is"."PLAN_ID"
where sr.reqcontent = '' OR sr.reqcontent IS NULL;

Once identified, reach out to the respective user to update the JQL query.