We have a high traffic site, with plenty of work done to make use of queued tracking and cron-based archiving. Now we’re hitting the point where transitions are giving us an Oops message later in the day.
I understand these are live queries, rather than archiving-based, and from the code, it looks like the queries themselves are governed by the ‘live_query_max_execution_time’ setting.
When I change this setting to 300, I can see the hint be included in the query to the database.
However, we still get an Oops message on the UI (behind the pop-up) after 30s, and no change to the foreground popup after any amount of time.
We implemented the suggested “ADD INDEX” commands, and that had the desired impact when viewing transitions from the Pages screen (ie via URLs), but it doesn’t help when viewing transitions from the Page Titles screen. In the DB, the query times are as long as before.
The SQL looks slightly different, so I guess an EXPLAIN will show a different index dependency list