Closed jeinwag closed 2 months ago
Sure, if you don't mind implementing this, that would be great. Additionally, could you please share details of your system, OS, and CRI-O installation so we can reproduce the issue?
Of course: OS: Ubuntu 24.04.01 Kubernetes: 1.30.4 CRI-O: 1.30.3
Kubernetes and CRI-O are deployed via kubespray v2.26.0
Will open a PR tomorrow.
Hi,
we're currently giving coroot a try. Unfortunately, it doesn't collect any metrics from any kubernetes pods (with CRI-O as the container runtime). I guess this is due to the inability to initialize the CRI-O client:
The location of the socket on the host is /run/crio/crio.sock while /var/run is just a symlink to /run, so /1/proc/root/var/run will point to the wrong location (/run of the container's filesystem).
Maybe some kind of retry logic like it does exist for different possible locations of the containerd.sock should be implemented?