CPU usage Concourse


Currently we have deployed concourse onto one of our VMs etc. and was wondering if I have deployed it correctly

As it keeps doing (img attached )even if its idle:

It used to build everything fine then suddenly just never finishes running its jobs etc. timeout etc.

Is there something I’m missing as this was fine then this suddenly occurred?

Concourse Version: v5.5.4

Help would be very much appreciated.

checking our concourse master which runs on an AWS t3a.small (dual core AMD 2.1GHz, 2GB RAM), and has 37 pipelines (each one associated with a git repo), and there’s typically one main concourse web process which bounces CPU usage up and down, usually about 10 to 15% but can hit 30% for a few seconds. We probably have about 10 to 15 people watching the web dashboard.

thus I would say your 60% CPU for one concourse web seems quite high, but that depends on how many pipelines you have and how many users you have watching the dashboard.

your load average suggests that there’s quite a few processes which are runnable but not being serviced, have you checked why they’re waiting - use “ps -efwl” and look at the WCHAN column and see if there’s a load stuck in futex or iowait.

Thanks for the reply however below is all I could find

For reference for our web we have around 20 pipelines btw

I should have mentioned that the t3a.medium I’m using as a concourse master is JUST the master. I have three dedicated concourse workers, which are m5a.xlarge and get quite busy!

Your worker nodes should be sized according to your own loading, I cannot help you there.