GAP Documentation
GitHub Toggle Dark/Light/Auto mode Toggle Dark/Light/Auto mode Toggle Dark/Light/Auto mode Back to homepage
  1. Let's get started!
  2. /
  3. Platform admin
  4. /
  5. Upgrades
  6. /
  7. Cluster Upgrade
Edit page

Cluster Upgrade

Cluster upgrade

  1. Check the Kubernetes changelog
  2. (for each staging, production, hq) Run kubepug to see any deprecated APIs in use.
  3. Announce the upgrade in the #infra-announcements Slack channel beforehand with the template:

⚠️ GAP <staging/production/hq>

Dear Users,

We’re starting a cluster upgrade on the <staging/production/hq> cluster. No issues are expected (if-stage-or-production:, but some Deployment Replicas Unavailable alerts might fire during this time.)

  1. Upgrade the control plane node:
    1. In Google Cloud Console upgrade the cluster control plane (GKE -> Cluster -> Details -> Version -> click Upgrade)
    2. Set the target version to the desired version
    3. Click Save Changes
  2. Upgrade the node pools (Node pool details -> Edit -> Node version) in the following order for <staging/production>:
    1. Cluster-components
    2. Whitelist-internal (Make sure the gap-staging-whitelist-internal-ip-* whitelist-internal fix IPs are assigned to nodes in GKE under VPC Network/IP addresses)
    3. Whitelist (Make sure the gap-staging-whitelist-ip-* whitelist fix IPs are assigned to nodes in GKE under VPC Network/IP addresses)
    4. Standard pool
    5. Baseline pool (on staging, you can upgrade Standard and Baseline pools together)
    6. Ingress
  3. Upgrade the node pools in the following order for :
    1. Cluster-components
    2. Baseline pool
  4. Update the kindest/node version here and here for the CI to validate the chart against the current kubernetes version
  5. Check the dashboards and alerts to make sure Alles ist in Ordnung.