Using Advanced Settings for custom fields, we can force a Jira single line text field into being a dimension with setting dimension=true
However the associated documentation warns: “Enable only for list-of-values custom field types and not for custom fields which have unique values for each issue”
In our case what is stored in field is peoples’ names e.g. “John Smith”, “Mary Jones” and we currently have about 2,400 name expect these names to grow only slowly over time. Note that these are not ‘user’ fields in Jira (they don’t have Jira accounts).
Having done this on a dev machine, the report capability is what we need but I’m trying to not ignore the warning.
What is there a good way of knowing whether or not this is a bad idea for performance reasons?
The only idea I have is to look at the cube size. With 27 custom fields and 2,443 issue, the cube size is 100.5 MB. After adding this field as a dimension (it was already being imported as a property), the cube size increased to 104.3 MB. This is a small cube and the increase is significant for such a small change (just 1 field). So I’m guessing this is okay for this cube/scenario.
What if the number of unique values was not 2,400 but more like 10,000 or 24,000? There is so much activity with Jira that an overall system performance test is difficult to do.
Thanks