Ask HN: Anything to improve DX of K8s kubectl?
I've been wondering if there is a simple fix for k8s unintuitive dx. A couple of things I find more difficult than should be:
- Deploy a pod, some controller doesn't agree. Finding out the controller is a bit more difficult (why aren't the controller logs aggregated and share a common saga id/key). I understand the evented nature of k8s with its controllers, but still it makes it very hard to trace and debug. - Why is it so hard to see _everything_ that is in a namespace. - Why isn't the YAML typed and so elaborate. Or why can't we prefer a typed JS json/json5 since the actual objects are JS and adhere to a spec? - why are there specs in specs in the templates and templates? - Why is it `kubctl (nothing here) logs` and `kubectl get pods` - Deleting a pod and patching the finalizers?
I know this complexity is somehow needed but it would be nice if the 80% would be easy and the 20% would be possible. When it works it is fine. But when it doesn't its an absolute pain.
Also 80% of my time goes in the same repetitive tasks. Ideas are welcome!
No comments yet.