Marked branch as 'abandoned'. We had a conversation here (adeuring, gmb, kfogel) and have come around to a solution that doesn't need a schema change. You're right: the fact that bug heat is calculated as a background task makes the difference; and gmb explained that current performance problems with the subscriber portlet are not due to database queries but to over-iteration after the query (so there is no need for us to try to address those problems here).
Marked branch as 'abandoned'. We had a conversation here (adeuring, gmb, kfogel) and have come around to a solution that doesn't need a schema change. You're right: the fact that bug heat is calculated as a background task makes the difference; and gmb explained that current performance problems with the subscriber portlet are not due to database queries but to over-iteration after the query (so there is no need for us to try to address those problems here).