Logging

The Actions Runner Controller (ARC) resources, which include the controller, listener, and runners, write logs to standard output (stdout). We recommend you implement a logging solution to collect and store these logs. Having logs available can help you or GitHub support with troubleshooting and debugging. For more information, seeLogging Architecture in the Kubernetes documentation.

Resources labels

Labels are added to the resources created by Actions Runner Controller, which include the controller, listener, and runner pods. You can use these labels to filter resources and to help with troubleshooting.

Controller pod

The following labels are applied to the controller pod.

app.kubernetes.io/component=controller-managerapp.kubernetes.io/instance=<controllerinstallationname>app.kubernetes.io/name=gha-runner-scale-set-controllerapp.kubernetes.io/part-of=gha-runner-scale-set-controllerapp.kubernetes.io/version=<chartversion>

Listener pod

The following labels are applied to listener pods.

actions.github.com/enterprise=# Will be populated if githubConfigUrl is an enterprise URLactions.github.com/organization=# Will be populated if githubConfigUrl is an organization URLactions.github.com/repository=# Will be populated if githubConfigUrl is a repository URLactions.github.com/scale-set-name=# Runners scale set nameactions.github.com/scale-set-namespace=# Runners namespaceapp.kubernetes.io/component=runner-scale-set-listenerapp.kubernetes.io/part-of=gha-runner-scale-setapp.kubernetes.io/version=# Chart version

Runner pod

The following labels are applied to runner pods.

actions-ephemeral-runner=# True | Falseactions.github.com/organization=# Will be populated if githubConfigUrl is an organization URLactions.github.com/scale-set-name=# Runners scale set nameactions.github.com/scale-set-namespace=# Runners namespaceapp.kubernetes.io/component=runnerapp.kubernetes.io/part-of=gha-runner-scale-setapp.kubernetes.io/version=# Chart version

Checking the logs of the controller and runner set listener

To check the logs of the controller pod, you can use the following command.

kubectl logs -n <CONTROLLER_NAMESPACE> -l app.kubernetes.io/name=gha-runner-scale-set-controller

kubectl logs -n <CONTROLLER_NAMESPACE> -l app.kubernetes.io/name=gha-runner-scale-set-controller

To check the logs of the runner set listener, you can use the following command.

kubectl logs -n <CONTROLLER_NAMESPACE> -l auto-scaling-runner-set-namespace=arc-systems -l auto-scaling-runner-set-name=arc-runner-set

kubectl logs -n <CONTROLLER_NAMESPACE> -l auto-scaling-runner-set-namespace=arc-systems -l auto-scaling-runner-set-name=arc-runner-set

Using the charts from themaster branch

We recommend you use the charts from the latest release instead of themaster branch. Themaster branch is highly unstable, and we cannot guarantee that the charts in themaster branch will work at any given time.

Troubleshooting the listener pod

If the controller pod is running, but the listener pod is not, inspect the logs of the controller first and see if there are any errors. If there are no errors and the runner set listener pod is still not running, ensure the controller pod has access to the Kubernetes API server in your cluster.

If you have a proxy configured or you're using a sidecar proxy that's automatically injected, such asIstio, ensure it's configured to allow traffic from the controller container (manager) to the Kubernetes API server.

If you have installed the autoscaling runner set, but the listener pod is not created, verify that thegithubConfigSecret you provided is correct and that thegithubConfigUrl you provided is accurate. SeeAuthenticating ARC to the GitHub API andDeploying runner scale sets with Actions Runner Controller for more information.

Runner pods are recreated after a canceled workflow run

Once a workflow run is canceled, the following events happen.

• The cancellation signal is sent to the runners directly.
• The runner application terminates, which also terminates the runner pods.
• On the next poll, the cancellation signal is received by the listener.

There might be a slight delay between when the runners receive the signal and when the listener receives the signal. When runner pods start terminating, the listener tries to bring up new runners to match the desired number of runners according to the state it's in. However, when the listener receives the cancellation signal, it will act to reduce the number of runners. Eventually the listener will scale back down to the desired number of runners. In the meantime, you may see extra runners.

Error:Name must have up to n characters

ARC uses the generated names of certain resources as labels for other resources. Because of this requirement, ARC limits resource names to 63 characters.

Because part of the resource name is defined by you, ARC imposes a limit on the number of characters you can use for the installation name and namespace.

Error: INSTALLATION FAILED: execution error at (gha-runner-scale-set/templates/autoscalingrunnerset.yaml:5:5): Name must have up to 45 charactersError: INSTALLATION FAILED: execution error at (gha-runner-scale-set/templates/autoscalingrunnerset.yaml:8:5): Namespace must have up to 63 characters

Error:Access to the path /home/runner/_work/_tool is denied

You may see this error if you're using Kubernetes mode with persistent volumes. This error occurs if the runner container is running with a non-root user and is causing a permissions mismatch with the mounted volume.

To fix this, you can do one of the following things.

• Use a volume type that supportssecurityContext.fsGroup.hostPath volumes do not support this property, whereaslocal volumes and other types of volumes do support it. Update thefsGroup of your runner pod to match the GID of the runner. You can do this by updating thegha-runner-scale-set helm chart values to include the following. ReplaceVERSION with the version of theactions-runner container image you want to use.

  YAML

  template:spec:securityContext:fsGroup:123containers:-name:runnerimage:ghcr.io/actions/actions-runner:latestcommand: ["/home/runner/run.sh"]

• If updating thesecurityContext of your runner pod is not a viable solution, you can work around the issue by usinginitContainers to change the mounted volume's ownership, as follows.

  YAML

  template:spec:initContainers:-name:kube-initimage:ghcr.io/actions/actions-runner:latestcommand: ["sudo","chown","-R","1001:123","/home/runner/_work"]volumeMounts:-name:workmountPath:/home/runner/_workcontainers:-name:runnerimage:ghcr.io/actions/actions-runner:latestcommand: ["/home/runner/run.sh"]

Error:failed to get access token for GitHub App auth: 401 Unauthorized

A401 Unauthorized error when attempting to obtain an access token for a GitHub App could be a result of a Network Time Protocol (NTP) drift. Ensure that your Kubernetes system is accurately syncing with an NTP server and that there isn't a significant time drift. There is more leeway if your system time is behind GitHub's time, but if the environment is more than a few seconds ahead, 401 errors will occur when using GitHub App.

Runner group limits

You can have a maximum of 10,000 self-hosted runners in one runner group. If this limit is reached, adding a new runner will not be possible.

Runner updates

Validate that your runner software version and/or custom runner image(s) in use are running the latest version.

For more information, seeSelf-hosted runners reference.

Legal notice

Portions have been adapted fromhttps://github.com/actions/actions-runner-controller/ under the Apache-2.0 license:

Copyright 2019 Moto IshizawaLicensed under the Apache License, Version 2.0 (the "License");you may not use this file except in compliance with the License.You may obtain a copy of the License at    http://www.apache.org/licenses/LICENSE-2.0Unless required by applicable law or agreed to in writing, softwaredistributed under the License is distributed on an "AS IS" BASIS,WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.See the License for the specific language governing permissions andlimitations under the License.