Project

General

Profile

Actions

Bug #20472

closed

Too many ProbesPerSecond kills the database

Added by Peter Amstutz 11 months ago. Updated 7 months ago.

Status:
Resolved
Priority:
Normal
Assigned To:
Category:
API
Story points:
-
Release relationship:
Auto

Description

I set the dispatcher ProbesPerSecond to 10 and it rapidly started to generate a request backlog on the API server. Seems like it is still fighting over the container table lock.

When I set it back to 5 per second, the backlog went away.

What the probes doing that causes it to slam the database like that?

Do container table updates need a separate limiter like the logs endpoint so they don't kill the API server?


Subtasks 1 (0 open1 closed)

Task #20478: Review 20472-priority-updateResolvedPeter Amstutz05/04/2023Actions
Actions

Also available in: Atom PDF