XXX: about, slides, weaveworks, brice
In order to conplete this tutorial, you will need a test kubernetes cluster. You have several options to quikcly get a development cluster to test out Gitops
- Run Minikube locally. This will require you to have Virtualbox installed. This is great if you have virtualbox installed, but the download is over 2GB. See the setup isntructions for Minikube.
- Use Google Kubernetes Engine. You'll need to have a google cloud instance set up to do this, and it will cost you a small amount to keep your cluster running. See the setup instructions for GKE
- Set up a cluster on DigitalOcean. This will also cost you a little to keep your cluster running. You can follow the setup instructions for Digital Ocean to do this.
You may have another way of getting a kubernetes cluster up and running. If that's the case, great! This tutorial should work with any valid kubernetes installation.
Before you begin, you should make sure you have configured the kubectl command to point to your cluster and can query basic information. for example, run
➤ kubectl get pods --all-namespaces
To get a listing of all the running pod on your cluster. If this command completes succesfully, you're ready to go onto the next step.
in order to control the operation of your cluster using gitops, you'll need to
- Add deploy keys with github repo
- Configure the deploy operator to point to your cloned repository, to the
deploy/kubernetesfolder
Now that the operator is set up to react to changes in your repository's deploy/kubernetes directory, we need to provide it with some configuration for it to synchronise.
Now that we know that the gitops operator is running in your cluster, let's check the control group
- look at number of running pod for service
- scale up deployment in the yaml
- commit
- watch the cluster for the change
- look at version of deployment
- add automation annotation
- watch to see the cluster update to latest version
- copy the monitoring folder yamls
- Look at the exposed service for grafana
- modify the dashboard script to add a new dashboard
- commit the changes.
Now you're done with this tutorial, make sure you delete your cluster so you don't incur running costs. See the