如何使用python API从GCP服务帐户以编程方式生成kubernetes配置

Question

I already found the way using gcloud CLI.

gcloud auth activate-service-account --key-file=serviceaccount.json
gcloud container clusters get-credentials $clusterName \
    --zone=$zone --project=$project
kubectl config view --minify --flatten

However, to eliminate dependency to gcloud cli, Is there any programmatic way to achieve a similar result as above? Preferably using API exposed in Google's python client library.

My expected result is a portable config file that can be passed to any kubectl --kubeconfig=... command.

---

update : I have found that the commands I showed above results in a kube config file that still depends on gcloud cli as auth helper, probably to automatically handle token expiration. So, any workarounds are welcome.

Answer 1

I wrote a shell script which basically does exactly what you are expecting.

https://gitlab.com/workshop21/open-source/rbac