we host the flowable engine in GCP and GCP comes with a nice feature called Query Insights. It keeps a dynamic list of the slowest queries and displays it on the dashboard.
At one point we started experiencing a significant slowdown of the platform and here is a screenshot of the Query Insights dashboard around that time:
As you can see the ones at the top are from the Flowable engine itself (as far as I can tell). Also, around that time the database wait lock time increased significantly.
Did anyone experience anything like that? How did you address this issue?
I’m not very sure about this, however, IOPS increases with the usage of Timer Events especially when you set time to minutes (PT1M). If you’ve got this event, think about whether this is needed and design a proper way to exit instead of infinite looping. Otherwise, I’m interested in other suggestions on this problem.
Looking at the top queries, it seems all to be updates.
How large is this table? What kind of database setup (hardware-wise) are you running this - it’s very odd that updates would be slow unless an index is missing (which could still be the case). Can you also paste the full SQL of the top queries?