A Kubernetes CLI wrapper with object-first syntax.

A simple alternative to kubectl.

Makes it easier to perform series of operations on k8s API objects.

If you manage kubernetes clusters, you probably very often find yourself issuing seies of commands similar to this:

$ kubectl get po
$ kubectl get po myapp-9bf86f6a7-k6rn3 -o wide
$ kubectl describe po myapp-9bf86f6a7-k6rn3
$ kubectl logs myapp-9bf86f6a7-k6rn3

The annoying bit in this flow is that you have to navigate backwards through your last command to change the verb. Here's how you would do the same with kobj:

$ kobj po
$ kobj po myapp-9bf86f6a7-k6rn3 get -o wide
$ kobj po myapp-9bf86f6a7-k6rn3 describe 
$ kobj po myapp-9bf86f6a7-k6rn3 logs 

That saves you a lot of keystrokes on the common tasks!

$ kobj <ObjectType> [get|list] [OPTIONS...]              - list objects of type ObjectType
$ kobj <ObjectType> <ObjectName> <Action> [OPTIONS...]   - perform Action on specified object

All the OPTIONS and Actions are the same as for kubectl. It really is just a simlpe wrapper around kubectl!

There are a few deviations from how kubectl does things currently:

• logs action still requires you to specify po/pod/pods as the ObjectType. This is for syntax consistency and to speed up the typical workflow.
• list is introduced as an alias to get

kobj is not written to replacekubectl, but rather to give you an alternative, object-focused perspective in situations where it's beneficial. So, as a rule of thumb:

• If you're troubleshooting a single object or several of the same type, you'll be more effective with kobj
• If you're exploring objects of various types, you'll be more effective with kubectl

Since it can be useful to alternate between kobj and kubectl perspective, kobj will always print the equivalent kubectl command at the bottom of the output, on stderr.