#What's Changed

[Fix] When the status check's first database query is slow, it now measures once more before reporting degraded performance. One-off delays, such as a fresh connection or the database waking up, no longer show up as incidents; a database that stays slow still does.