Code Monkey home page Code Monkey logo

building-platforms's Introduction

Building self serving platforms

The State of DevOps report 2023 from Puppet revealed some interesting developments in the evolution of devops as a practice. A significant finding was the emergence of the plaform or more specifically platform engineering. Platform engineering can be generally considered as the development of a product aware infrastructure that provisions self service capbilities. Platform teams continually adapt and evolve those services and enable stream aligned teams to access those capabilities in a frictionless and automated way. Teams consuming the platform are also aligned to the organisations value streams. The business outcome tends towards increased productivity through standardised practices, less duplication of work, better security and ultimately, a better quality product.

This blog demonstrates with examples how a self-serving encryption service may be incorporated as a platform service within your organisation or enterprise. The design goal is to address varied governance concerns across business units within the organsation through the imposition of boundaries that isolate vaults from one another whilst remaining within the same overall deployment. By creating low resistance interaction between platform and stream aligned teams, the cognitive load of both is reduced therby permitting teams to focus on what matters most, delivering value.

Hashicorp Vault is the use case. The integration section below will show how it can be integrated into OpenShift Container Platform from Red Hat to provide encryption as a service to your workloads. The examples in this article are for demonstrated purposes only and are in no way assuring relevance for any context. They should however be sufficiently adaptable to meet many businss scenarios.

Enterprise Vault introduces a feature called namespaces, they provide a means to create vaults within vault, each having their own encryption engines, storage backends and authentication methods. We are going to use this feature to create a number of namespaces. However, the OCP namespace in Vault, dedicated to workloads running in OpenShift is the sole focus. The others are purely illustrative, to demonstrate what a self serving capability can achieve.

  1. OCP - A namespace dedicated to workloads running in OpenShift.
  2. Users - A namespace dedicated to users to store their own key value secrets. They will use the CubbyHole facility and no storage backends. This namespace would probably integrate with an enterprises' internal Identity Management Provider (IDP).
  3. Legal - A namespace dedicated to users within a legal department. They could use a different or the same IDP as the User namespace.
  4. Admin - The default namespace. Use this just for the root token to be stored and accessed as a break glass. No storage backends or auth methods are configured here therby reducing attack vectors and improving the overall security posture.

A basic nginx webserver will be used as an example deployment to demonstrate the integration with Vault. The app will use templates to create html files that can be mounted as files and made availble to the webserver. All the required files are available in this repository. Please fork the repository to make use of them.

A more realistic use case for using the template approach would be to render an application properties file and mount it to a path where a framework [like Spring Boot] would expect to find it. That's a little heavy weight for this purpose but the idea of rendering web pages to a location where a server is expecting to find them demonstrates the approach just fine.

To follow this you will need to the following:

  • A Hashicorp Vault Enterprise cloud account. Sign up for a free trial and receive $50 credit in your account. You can use this to install a Development Small cluster. The budget should last around two months depending on your region and usage.
  • An OpenShift account with sufficient permissions to install the agent and ClusterRoleBindings

Vault CLI config

(i) Ensure that you have the Vault CLI installed and that you connect to the Vault server. Set up the following environment variables.

export VAULT_ADDR="https://your-vault-address:8200"
export VAULT_TOKEN="the-token"

(ii) Create a namespace in Vault and set an environment variable to use the name space for subsequent commands issued using the Vault CLI.

$ vault namespace create ocp   
$ export VAULT_NAMESPACE="admin/ocp" 

Let's start!

The big picture

The big picture

Install the vault agent in OpenShift Install the Vault Agent in OpenShift

Refer to the install agent guide in this repository

Setting up the Kubernetes Auth method Setting up the Kubernetes Auth method

(i) Enable the Kubernetes Auth engine in Vault. Use the Vault CLI.

$ vault auth enable kubernetes

(ii) Obtain the JWT token fom the Vault ServiceAccount. Use the oc command.

$ oc describe  sa vault -n hashicorp-vault         
Name:                vault
Namespace:           hashicorp-vault
Labels:              app.kubernetes.io/instance=vault
                     app.kubernetes.io/managed-by=Helm
                     app.kubernetes.io/name=vault
                     helm.sh/chart=vault-0.22.0
Annotations:         meta.helm.sh/release-name: vault
                     meta.helm.sh/release-namespace: hashicorp-vault
Image pull secrets:  vault-dockercfg-flp6c
Mountable secrets:   vault-token-x6hrv
                     vault-dockercfg-flp6c
Tokens:              vault-token-4ltc7
                     vault-token-x6hrv
Events:              <none>

(iii) Place the JWT token for the secret vault-token-x6hrv into an environment variable. Use the oc command.

$ export JWT_TOKEN=$(oc get secret vault-token-x6hrv -n hashicorp-vault -o jsonpath='{.data.token}' | base64 -d)

(iv) Add the Kubernetes API CA and the Kubernetes API endpoint to some environment variables. Use the oc command.

$ export KUBE_CA_CRT=$(oc get cm kube-root-ca.crt -n openshift-config -o jsonpath='{.data.ca\.crt}') 
$ export KUBE_HOST=$(oc config view --raw --minify --flatten --output 'jsonpath={.clusters[].cluster.server}')

(v) Create the auth configuration within Vault. Use the Vault CLI

$ vault write auth/kubernetes/config \
     token_reviewer_jwt="$JWT_TOKEN" \
     kubernetes_host="$KUBE_HOST" \
     kubernetes_ca_cert="$KUBE_CA_CRT" 

(vi) Verify the config. Use the Vault CLI

$ vault read auth/kubernetes/config

Creating secrets Create the secrets

(i) Enable the kv2 secrets engine. Use the Vault CLI

$ vault secrets enable -path=secrets kv-v2

(ii) Write two secrets to the vault. These will represent content for our HTML pages. Use the Vault CLI.

$ vault kv put secrets/my-app/dev/contact title="Contact us" telephone="+(44) 555-234545" 
$ vault kv put secrets/my-app/dev/index title="Home page" password="The secret from hashi vault" 

(iii) Verify the secrets. Use the Vault CLI.

$ vault kv get secrets/my-app/dev/contact 
$ vault kv get secrets/my-app/dev/index 

Create a policy Create a policy

(i) Create a policy that defines access to the secrets. Use the Vault CLI

$ vault policy write my-app-secrets - <<EOF
path "secrets/data/my-app/dev/\*" {
  capabilities = ["read" , "list"]
}
EOF

Create a role Create a role

(i) A role associates a ServiceAccount and NameSpaces to a policy. Use the Vault CLI.

$ vault write auth/kubernetes/role/my-app \
    bound_service_account_names=hello-world-app \
    bound_service_account_namespaces=hello-world \
    policies=my-app-secrets \
    ttl=24h

Create a role OpenShift Deployment

(i) Create templates to be rendered into html files containing the secrets from Vault. Use a text editor for this.

(ii) Create a secret to contain those files within Kubernetes. The files contact.html and index.html in this repository can be used. Use the oc command.

$ oc create secret generic vault-agent-template-web-pages --from-file=index.html.ctmpl=./index.html --from-file=contact.html.ctmpl=./contact.html \
     --dry-run=client -o yaml > vault-agent-template-web-pages.yaml

(iii) Apply the Vault annotations to the OpenShift deployment. Refer to deployment.yaml in this repository.

annotations: 
    vault.hashicorp.com/namespace: admin/ocp
    vault.hashicorp.com/agent-inject: 'true'
    vault.hashicorp.com/agent-pre-populate-only: 'true'
    vault.hashicorp.com/auth-path: 'auth/kubernetes'
    vault.hashicorp.com/role: 'my-app'    
    # this is a secret stored in vault 
    vault.hashicorp.com/agent-inject-secret-index.html: 'secrets/my-app/dev/index'
    vault.hashicorp.com/agent-inject-secret-contact.html: 'secrets/my-app/dev/contact'
    # this path to be rendered 
    vault.hashicorp.com/secret-volume-path-index.html: /usr/share/nginx/html
    vault.hashicorp.com/secret-volume-path-contact.html: /usr/share/nginx/html
    # this is the k8s secret that contains the templates which the agent combines using the vault secrets
    vault.hashicorp.com/agent-extra-secret: vault-agent-template-web-pages 
    # this tells vault where to load the template so it can get rendered 
    vault.hashicorp.com/agent-inject-template-file-contact.html: /vault/custom/contact.html.ctmpl
    vault.hashicorp.com/agent-inject-template-file-index.html: /vault/custom/index.html.ctmpl


building-platforms's People

Contributors

spiro-redhat avatar

Stargazers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.