rakkess

README

rakkess

Review Access - kubectl plugin to show an access matrix for server resources

Intro

Have you ever wondered what access rights you have on a provided kubernetes cluster? For single resources you can use kubectl auth can-i list deployments, but maybe you are looking for a complete overview? This is what rakkess is for. It lists access rights for the current user and all server resources, similar to kubectl auth can-i --list.

It is also useful to find out who may interact with some server resource. Check out the sub-command rakkess resource below.

Demo

Examples

Show access for all resources

• ... at cluster scope

  rakkess
  

• ... in some namespace

  rakkess --namespace default
  

• ... with verbs

  rakkess --verbs get,delete,watch,patch
  

• ... for another user

  rakkess --as other-user
  

• ... for another service-account

  rakkess --sa kube-system:namespace-controller
  

• ... and combine with common kubectl parameters

  KUBECONFIG=otherconfig rakkess --context other-context
  

Show subjects with access to a given resource¹

• ...globally in all namespaces (only considers ClusterRoleBindings)

  rakkess resource configmaps
  

• ...in a given namespace (considers RoleBindings and ClusterRoleBindings)

  rakkess resource configmaps -n default
  

• ...with shorthand notation

  rakkess r cm   # same as rakkess resource configmaps
  

• .. with custom verbs

  rakkess r cm --verbs get,delete,watch,patch
  

Name-restricted roles

Some roles only apply to resources with a specific name. To review such configurations, provide the resource name as additional argument. For example, show access rights for the ConfigMap called ingress-controller-leader-nginx in namespace ingress-nginx (note the subtle difference for nginx-ingress-serviceaccount to the previous example):

As rakkess resource needs to query Roles, ClusterRoles, and their bindings, it usually requires administrative cluster access.

Also see Usage.

Installation

There are several ways to install rakkess. The recommended installation method is via krew.

Via krew

Krew is a kubectl plugin manager. If you have not yet installed krew, get it at https://github.com/kubernetes-sigs/krew. Then installation is as simple as

kubectl krew install access-matrix

The plugin will be available as kubectl access-matrix, see doc/USAGE for further details.

Binaries

When using the binaries for installation, also have a look at doc/USAGE.

Linux

curl -LO https://github.com/corneliusweig/rakkess/releases/download/v0.5.0/rakkess-amd64-linux.tar.gz \
  && tar xf rakkess-amd64-linux.tar.gz rakkess-amd64-linux \
  && chmod +x rakkess-amd64-linux \
  && mv -i rakkess-amd64-linux $GOPATH/bin/rakkess

OSX

curl -LO https://github.com/corneliusweig/rakkess/releases/download/v0.5.0/rakkess-amd64-darwin.tar.gz \
  && tar xf rakkess-amd64-darwin.tar.gz rakkess-amd64-darwin \
  && chmod +x rakkess-amd64-darwin \
  && mv -i rakkess-amd64-darwin $GOPATH/bin/rakkess

Windows

https://github.com/corneliusweig/rakkess/releases/download/v0.5.0/rakkess-windows-amd64.zip

From source

Build on host

Requirements:

• go 1.16 or newer
• GNU make
• git

Compiling:

export PLATFORMS=$(go env GOOS)
make all   # binaries will be placed in out/

Build in docker

Requirements:

• docker

Compiling:

mkdir rakkess && chdir rakkess
curl -Lo Dockerfile https://raw.githubusercontent.com/corneliusweig/rakkess/master/Dockerfile
docker build . -t rakkess-builder
docker run --rm -v $PWD:/go/bin/ --env PLATFORMS=$(go env GOOS) rakkess
docker rmi rakkess-builder

Binaries will be placed in the current directory.

Users

What are others saying about rakkess?
“Well, that looks handy! rakkess, a kubectl plugin to show an access matrix for all available resources.” – @mhausenblas
“that's indeed pretty helpful. rakkess --as system:serviceaccount:my-ns:my-sa -n my-ns prints the access matrix of a service account in a namespace” – @fakod
“THE BOMB. Love it.” – @ralph_squillace
“This made my day. Well, not actually today but I definitively will use it a lot.” – @Soukron

---

[1]: This mode was inspired by kubectl-who-can