gitlab-runner.yaml

#
# References:
# - https://docs.gitlab.com/12.4/runner/install/kubernetes.html
# - https://docs.gitlab.com/runner/executors/kubernetes.html#workflow
# - https://gitlab.com/gitlab-org/charts/gitlab-runner/blob/master/values.yaml

repositories:
## Add the Gitlab Helm repository
- name: gitlab
  url: "https://charts.gitlab.io"

releases:
- name: "gitlab-runner"
  namespace: "gitlab-runner"
  labels:
    chart: "gitlab-runner"
    repo: "gitlab"
    component: "gitlab-runner"
    namespace: "gitlab-runner"
    vendor: "gitlab"
    default: "false"
  chart: "gitlab/gitlab-runner"
  # version: "0.10.1"
  version: "0.15.0"
  installed: {{ env "GITLAB_RUNNER_INSTALLED" | default "false" }}
  wait: true
  values:
    - runnerRegistrationToken: {{ env "GITLAB_RUNNER_TOKEN" | default "" }}
      gitlabUrl: {{ env "GITLAB_URL" | default "https://gitlab.com" }}
      runners:
        ## Run all containers with the privileged flag enabled
        ## This will allow the docker:dind image to run if you need to run Docker
        ## commands. Please read the docs before turning this on:
        ## ref: https://docs.gitlab.com/runner/executors/kubernetes.html#using-docker-dind
        ##
        privileged: true
        # 
        # secret: gitlab-runner
        ## Gitlab Runner will assign nodes with given labels
        ## Always use a dedicated node group for gitlab runner jobs to avoid hurting other pods in the same nodes
        {{- if eq (env "KUBE_CLUSTER_PROVIDER" | default "eks" ) "eks" }}
        nodeSelector:
          purpose: gitlab-runner
        {{- end }}
        {{- if eq (env "KUBE_CLUSTER_PROVIDER" ) "doks" }}
        nodeSelector:
          doks.digitalocean.com/node-pool: gitlab-runner
        {{- end }}
        builds:
          limits:
            memory: 1024Mi
            cpu: 500m
          requests:
            memory: 64Mi
            cpu: 100m
      rbac:
        create: true
      resources:
        limits:
          memory: 512Mi
          cpu: 500m
        requests:
          memory: 64Mi
          cpu: 100m
      ## Gitlab Runner should not be placed to Spot instances, use always regular nodes, because
      ## gitlab manager is responsible to handle placement of jobs and does not require a lot of resources
      affinity:
        requiredDuringSchedulingIgnoredDuringExecution:
          nodeSelectorTerms:
          - matchExpressions:
            - key: purpose
              operator: NotIn
              values:
              - gitlab-runner
        # preferredDuringSchedulingIgnoredDuringExecution:
        # - weight: 1
        #   preference:
        #     matchExpressions:
        #     - key: purpose
        #       operator: NotIn
        #       values:
        #       - gitlab-runner