What are your experiences with db performance tuning? I'm thinking about which indices we use internally and which of them are worthy to create in default installations. Similar situation is with partitioning, clustering and few other things we can do. But most of these are probably beneficial only with big deployments (Few versions ago I've started docs about db which now contains only partitioning). Is there something you would like to see related to db maintenance/performance?

/ Followup of https://pagure.io/koji/pull-request/2419 /

--

            Tomas Kopecek <tkopecek@redhat.com>
            RHEL Build Development, RedHat