agentk

"k"

TLDR;

Install the "k" by either doing:

pip install agentk

(Yes, ^^ it is written in python and your OS needs to have recent version 2 or 3)

or copying it in some bin folder on your PATH and running pip install -r requirements.txt

---

"A person is smart. People are dumb, panicky, dangerous animals, and you know it." -- Agent K

"AGENT" K is a complete minimalistic kubectl "doner"-wrap

Obviously, as a short-hand wrapper, k can do everything kubectl already can, but it is (a) shorter and (b) adds few tricks like merging configs and switching contexts .. (k) feeds back to the kubectl command-line those args which it does not want to intercept or handle.

Usage

The following is equivalent:

kubectl get pods --all-namespaces
k get pods -A
k p -A

Switching context

Argument-free invocation prompts for context switch options between multiple cluster contexts found in ~/.kube/config:

k

Switching namespaces

One can change the default namespace on the currently active context (namespace key in ~/.kube/config) using either of two equivalent commands:

kubectl config set-context $(kubectl config current-context) --namespace foo
k sn foo

The last command is a k shortcut.

Shortcuts to get resources

You can find the full list of shortcuts defined as the dictionary inside the k script. In particular that would be:

# resource
"ev": "event",
"ep": "endpoints",
"p": "pod",
"s": "service",
"v": "volume",
"n": "node",
"dp": "deployment",
"st": "statefulset",
"in": "ingress",
"ns": "namespace",

At the end of the list there are one letter action-shortcuts:

# actions
"c": "create",
"a": "apply",
"d": "delete",

This means that the following is equivalent:

kubectl apply -f <foo-k8s-manifest.yaml>
k a -f <foo-k8s-manifest.yaml>

Develop

To remind, you can do pip install -e . in order to utilize developer mode.

Installation in the cloud

If you work with kubectl without a privileged or super-user access, for example inside a corporate network or in a cloud-shell (but you still have access to python), then your installation will look like:

pip install --user agentk

This will install the script in your local $HOME folder.

Don't forget to append your ~/.bashrc or ~/bash_profile or other shell-rc file with:

export PATH="$HOME/.local/bin:$PATH"

Command completion

You can put this into your .bashrc to get alias and auto completion for k similar as for kubectl:

source <(kubectl completion bash | sed s/kubectl/k/g)

Similar works well for zsh.