Tutorial part 4: Cluster with remote GitOps repo and prod infrastructure
This fourth part of the tutorial continues from the end of Part 3. If not done already, follow the Installation instructions, make sure you have cleaned up any deployments you made in Part 1, and set up the cluster as described in Parts 2 and 3.
You should have a cluster deployed to your node machines with ArgoCD monitoring your private external repo for changes to configuration, and SSl certificates being served by the Let's Encrypt staging endpoint.
If not done already, follow the Installation instructions.
Checks prior to launching prod
- Remove the default ArgoCD admin password and set your own. This is possible through the ArgoCD UI.
- Ensure the Traefik dashboard has a password.
- Optional: Set up backups and snapshots in Longhorn. See the Longhorn docs.
Change all environments to prod
Ensure the following ArgoCD marker files are renamed, and any sibling environment marker files are renamed to config.json.ignore:
infrastructure/argocd/envs/prod/config.jsoninfrastructure/cert-manager/prod/config.jsoninfrastructure/kubevip/envs/prod/config.jsoninfrastructure/longhorn/envs/prod/config.jsoninfrastructure/metallb/envs/prod/config.jsoninfrastructure/sealed-secrets/envs/prod/config.jsoninfrastructure/traefik/envs/stage/config.json
As well as the test tenant applications:
tenants/test-ingress/envs/prod/config.jsontenants/test-lb/envs/prod/config.json
Commit and push the changes to the private GitOps repo.
It should migrate within 3 minutes, or manually sync the applications in the ArgoCD UI. You should no longer see certificate errors when visiting https:// page hosted by the cluster.
Congratulations!
You've deployed the cluster to production. Now you can deploy your own tenant workloads to the cluster.
Created: June 13, 2024 16:06:52