In server profiling log, I’m not getting SlowQuery tags even after hours of run.
--long-database-query-threshold (Timespan): The threshold for a
database query to be considered slow
If any query takes longer than this threshold, it will be logged
with a warning as a slow query. Set this value to less than 0
seconds to disable the monitor.
* default value: 30m
I’m actually on 2.3.1.0_beta_2026-07-28, but this might be its only announce.
It “looks” like the feature continued in Canary, but didn’t sneak into 2.3 Stable.
I’m looking in a live log at Warning level, and a profiling log of server activities.
EDIT 1:
It sometimes works, sometimes not. I saw it in 2.3.0.108_canary_2026-07-20.
Had to be dealt with awkwardly. I set 1s, and production backup warned some. 1m at least isn’t killed by job edit, but I set that, and in 5 minutes got no warning
from the “Waiting for upload to finish” test, but I’m pretty sure it’s stuck in SQLite.
Maybe it doesn’t like that that query also escapes the log, or its timing is nested?