Deploying actions-runner-controller on K3S instance

Questions and Answers Deployment
,
1

Hello @Qovery_Team, @Pierre_Mavro
I am trying to setup a self-hosted github actions runner on our K3S single node EC2. Will that create any problems with Qovery control plane?

I am getting an error that kubernetes cluster is unreachable, but kubetctl shows all pods are running (qovery, cert-manager,kube-system) (error message below). Any ideas?

sudo helm install --set authSecret.github_token=“” --set authSecret.create=true --wait --namespace actions-runner-system --create-namespace actions-runner-controller actions-runner-controller/actions-runner-controller

Error: INSTALLATION FAILED: Kubernetes cluster unreachable: Get “http://localhost:8080/version”: dial tcp 127.0.0.1:8080: connect: connection refused

2

Hi @sama213 ,

I don’t know where you run the helm command from, but it looks like 127.0.0.1 is not the right target. Did you get the kubeconfig and setup it up accordingly in the doc?

Also, Github action runner will not interfere with Qovery control panel, but I advise you to set limits and requests (CPU & memory) correctly on all deployed objects or you’ll certainly crash your EC2 instance.

Just a few words from my POV, running Github runner controller doesn’t make much more sense than that on a single machine (depending on the size, of course), because you’ll be quickly limited (running multiple pipelines in parallel). EKS cluster with autoscaling is much more adapted from my POV. However, if it’s just for a quick try, then it makes sense.