Add script for starting kind for local development.

[smuda]

Closes #1052
A script for starting kind from Makefile, to be used during development.

[pb82]

@NissesSenap since you're using kind, would you mind giving this a try?

[weisdd]

@smuda Thanks for the PR! I'd think it's better to maintain all the manifests in a separate folder (whether it's a reference to existing deploy examples or new ones). With that, we'd get syntax highlighting, code autocompletion and other things. WDYT?

[smuda]

Sure! These manifests was retrieved in the same manner that the DEVELOPMENT.md instructed, but I agree. I'll extract the manifests to a separate folder.

[NissesSenap]

So after fixing missing defnition of KUBECONFIG when you create your kind cluster I get the following issue:

###############################
# 1. Start kind cluster       #
###############################
Creating cluster "kind" ...
 ✓ Ensuring node image (kindest/node:v1.25.3) 🖼
 ✓ Preparing nodes 📦  
 ✓ Writing configuration 📜 
 ✓ Starting control-plane 🕹️ 
 ✓ Installing CNI 🔌 
 ✓ Installing StorageClass 💾 
 ✓ Waiting ≤ 2m0s for control-plane = Ready ⏳ 
 • Ready after 19s 💚
Set kubectl context to "kind-kind"
You can now use your cluster with:

kubectl cluster-info --context kind-kind --kubeconfig /home/edvin/.kube/kind-grafana-operator

Not sure what to do next? 😅  Check out https://kind.sigs.k8s.io/docs/user/quick-start/
The connection to the server localhost:8080 was refused - did you specify the right host or port?
make: *** [Makefile:272: start-kind] Error 1

This is due to you don't specify the custom kubeconfig file wihtin your command

kubectl get pods --context kind-kind --kubeconfig /home/edvin/.kube/kind-grafana-operator

Please take a second look at at the script and I will be happy to merge.
Thanks allot @smuda

[NissesSenap]

Maybe it would be good to have a none standard name like grafana-kind or something like that.
This way we don't remove someone's cluster that they do other things on.

[smuda]

I've set the default name to "kind-grafana".

[NissesSenap]

Sure! These manifests was retrieved in the same manner that the DEVELOPMENT.md instructed, but I agree. I'll extract the manifests to a separate folder.

Agree, my basic thinking was that people wanted something to copy and paste quickly + I haden't solved the file import feature in the UI that I found later when I did this document.

I think a separate folder with a kustomize file in it would be nice.

[smuda]

I've moved resources to separate folders and added very simple kustomization.yaml.

With that, I think it's time for another round of reviews. :-)

Edit: I've also rebased on master.

[NissesSenap]

You need a final fix in the makefile.
It would also be nice if you can mention the usage of this script in CONTRIBUTING.md.
Other than that, it looks good.

[NissesSenap]

You are not overwriting the default kubeconfig since you are doing make install

[smuda]

Since I changed the default KUBECONFIG in the make file, I also had to explicitly specify the KUBECONFIG in the github e2e action.

[NissesSenap]

Sorry for the delay in looking at the PR.

Minor comment.

We should probably add something kind export kubeconfig -n kind-grafana or something like that to explain how to reach the instance after it's up and running.

[NissesSenap]

Should probably still have something around, make run . Right now when you start the cluster nothing will happen.

[smuda]

I added a small instruction on how to configure KUBECONFIG to be able to access the cluster.

Yes, once the cluster is up and running, nothing will happen. The idea was to create a cluster which is usable for development, much like the previous instructions in the CONTRIBUTING.md file. I've added an instruction pointing to make run to run the operator locally and changed the client.preferIngress=true so it works.

[NissesSenap]

LGTM, thanks @smuda for this. I think it will help a lot of other developers to kick-start their workflow.