The developer at Mystique Unicorn are interested in building their application using Kubernetes.They are looking for some help to secure their pods?
As part of a defense-in-depth strategy - You have two options to secure your custer
Kubernetes Pod Security Policy are cluster wide resources that control security sensitive attributes of pod specification and are a mechanism to harden the security posture of your Kubernetes workloads. It provides a method to enforce best practices around minimizing container runtime privileges, including not running as the root user, not sharing the host nodeβs process or network space, not being able to access the host filesystem, enforcing SELinux, and other options. Most cluster workloads will not need special permissions. By forcing containers to use the least-required privilege, their potential for exploitation can be minimized.
As a quick example, using PSPs you can:
- Prevent privileged pods from starting and control privilege escalation.
- Restrict access to the host namespaces, network and filesystem the pod can access.
- Restrict the users/groups a pod can run as.
- Limit the volumes a pod can access.
- Restrict other parameters like runtime profiles or read-only root file systems.
In this article, weβll show you how to harden your Kubernetes security posture by enabling a simple Pod Security Policy in your EKS environment.
Amazon EKS clusters with Kubernetes version 1.13
3 and higher have a default pod security policy named eks.privileged
. This policy has no restriction on what kind of pod can be accepted into the system, which is equivalent to running Kubernetes with the PodSecurityPolicy controller disabled.
Without a mechanism to restrict the pod spec privileges, the attacker can do anything the docker command could: running a privileged container, using node resources, etc.
For a quick proof of concept, you can run this script(stacks/k8s_utils/security_exploits/become_root_on_node.sh
)(DO NOT RUN THIS ON A PRODUCTION CLUSTER!).
# sh ./become_root_on_node.sh
[root@sudo--ip-10-10-1-67 /]# whoami
root
[root@sudo--ip-10-10-1-67 /]# hostname
sudo--ip-10-10-1-67.us-east-2.compute.internal
[root@sudo--ip-10-10-1-67 /]# ls -l /var/log/pods/
total 0
drwxr-xr-x 3 root root 21 Jun 15 10:58 default_miztiik-sudo_f4bf5609-80ce-4e75-8c12-714042dc4f74
drwxr-xr-x 3 root root 21 Jun 14 14:30 default_privileged_6155be2f-0c18-41ad-aaf5-c301f1495326
drwxr-xr-x 3 root root 27 Jun 13 21:03 default_ssm-installer-bk6rf_4adc663c-e22b-448f-bbbc-ab14e15c0a2f
drwxr-xr-x 2 root root 6 Jun 14 14:40 default_test_3aa9c15e-e1b1-4dbe-9774-cff36907d926
drwxr-xr-x 3 root root 34 Jun 15 10:52 kubernetes-dashboard_kubernetes-dashboard-5d9c5bb99d-stc6g_9a7781f1-cf03-4583-a7c8-20efe20c963f
drwxr-xr-x 4 root root 46 Jun 13 21:02 kube-system_aws-node-ccj5z_b3f5beff-5c73-4416-b801-1fd75562db2a
drwxr-xr-x 3 root root 24 Jun 13 21:02 kube-system_kube-proxy-rws4c_5b3f059e-b6e2-4eca-b6ef-8ce81368b71b
This is scary and highly undesirable.
Having said all the scary stuff, Are PSPs the way to secure you pods now and in the future. The kubernetes community thinks otherwise. PodSecurityPolicy is deprecated6,7 as of Kubernetes v1.21
, and will be removed in v1.25
. If that is being deprecated, what do we need to use to secure our cluster, The community has developed a Kubernetes Enhancement Proposal (KEP 2579) and a prototype for a new feature, currently being called by the temporary name "PSP Replacement Policy." It is being targeting for an Alpha release in Kubernetes 1.22
. Until then we can continue using the current PSPs
Let me show a very simple security policy, that will stop anyone from running a privileged pod. You can learn more on pod security best practices here9. We will use a simple AWS EKS cluster with couple of nodes to setup PSP.
-
This demo, instructions, scripts and cloudformation template is designed to be run in
us-east-1
. With few modifications you can try it out in other regions as well(Not covered here).- π AWS CLI Installed & Configured - Get help here
- π AWS CDK Installed & Configured - Get help here
- π Python Packages, Change the below commands to suit your OS, the following is written for amzn linux 2
- Python3 -
yum install -y python3
- Python Pip -
yum install -y python-pip
- Virtualenv -
pip3 install virtualenv
- Python3 -
-
-
Get the application code
git clone https://github.com/miztiik/eks-security-with-psp cd eks-security-with-psp
-
-
We will use
cdk
to make our deployments easier. Lets go ahead and install the necessary components.# You should have npm pre-installed # If you DONT have cdk installed npm install -g aws-cdk # Make sure you in root directory python3 -m venv .venv source .venv/bin/activate pip3 install -r requirements.txt
The very first time you deploy an AWS CDK app into an environment (account/region), youβll need to install a
bootstrap stack
, Otherwise just go ahead and deploy usingcdk deploy
.cdk bootstrap cdk ls # Follow on screen prompts
You should see an output of the available stacks,
eks-cluster-vpc-stack eks-cluster-stack ssm-agent-installer-daemonset-stack
-
Let us walk through each of the stacks,
-
Stack: eks-cluster-vpc-stack To host our EKS cluster we need a custom VPC. This stack will build a multi-az VPC with the following attributes,
- VPC:
- 2-AZ Subnets with Public, Private and Isolated Subnets.
- 1 NAT GW for internet access from private subnets
Initiate the deployment with the following command,
cdk deploy eks-cluster-vpc-stack
After successfully deploying the stack, Check the
Outputs
section of the stack. - VPC:
-
Stack: eks-cluster-stack As we are starting out a new cluster, we will use most default. No logging is configured or any add-ons. The cluster will have the following attributes,
- The control pane is launched with public access. i.e the cluster can be access without a bastion host
c_admin
IAM role added to aws-auth configMap to administer the cluster from CLI.- One managed EC2 node group created from a launch template
- It create two
t3.medium
instances runningAmazon Linux 2
- Auto-scaling Group with
2
desired instances. - The nodes will have a node role attached to them with
AmazonSSMManagedInstanceCore
permissions
- It create two
In this demo, let us launch the EKS cluster in a custom VPC using AWS CDK. Initiate the deployment with the following command,
cdk deploy eks-cluster-stack
After successfully deploying the stack, Check the
Outputs
section of the stack. You will find the*ConfigCommand*
that allows yous to interact with your cluster usingkubectl
-
Stack: ssm-agent-installer-daemonset-stack This EKS AMI used in this stack does not include the AWS SSM Agent out of the box. If we ever want to patch or run something remotely on our EKS nodes, this agent is really helpful to automate those tasks. We will deploy a daemonset that will run exactly once? on each node using a cron entry injection that deletes itself after successful execution. If you are interested take a look at the deamonset manifest here
stacks/back_end/eks_cluster_stacks/eks_ssm_daemonset_stack/eks_ssm_daemonset_stack.py
. This is inspired by AWS guidance.Initiate the deployment with the following command,
cdk deploy ssm-agent-installer-daemonset-stack
After successfully deploying the stack, You can use connect to the worker nodes instance using SSM Session Manager.
-
-
We are all set with our cluster to configure a custom pod security policy. NOTE: When you delete the default policy, no pods can be created on the cluster, except those that meet the security context in your new namespace.
-
Creating Custom Pod Security Policy in Kubernetes
We will need the following actions to setup a new PSP
- Create a new namespace:
miztiik-psp-restrictive
- Associate a service account with this namespace
sa-miztiik-dev
- Create new PSP
miztiik-secured
- Create a cluster role in this namespace
- Create role binding the role with the service account and namespace
- Apply new PSP to namespace
- Delete old/default PSP
- Try to launch privileged pod
- Create a new namespace:
-
Create Namespace and Service Account:
# Create a new namespace kubectl create ns miztiik-psp-restrictive # Create Service Account kubectl -n miztiik-psp-restrictive create sa sa-miztiik-dev # Create a cluster level Role for the service account to use kubectl -n miztiik-psp-restrictive create rolebinding miztiik-psp-admin \ --clusterrole=edit \ --serviceaccount=miztiik-psp-restrictive:sa-miztiik-dev
-
Create Pod Security Policy: The PSP policy in this directory
stacks/k8s_utils/pod_security_policies/secure_psp_policy.yaml
, save this assecure_psp_policy.yaml
apiVersion: policy/v1beta1 kind: PodSecurityPolicy metadata: name: miztiik-secured spec: privileged: false # DO NOT ALLOW privileged pods! seLinux: rule: RunAsAny supplementalGroups: rule: RunAsAny runAsUser: rule: RunAsAny fsGroup: rule: RunAsAny volumes: - "*"
kubectl -n miztiik-psp-restrictive apply -f secure_psp_policy.yaml
You can confirm if the policy has been applied correctly,
kubectl get psp -n miztiik-psp-restrictive
Expected Output,
NAME PRIV CAPS SELINUX RUNASUSER FSGROUP SUPGROUP READONLYROOTFS VOLUMES miztiik-secured false RunAsAny RunAsAny RunAsAny RunAsAny false *
-
Create Role & Rolebinding:
# Create Role kubectl -n miztiik-psp-restrictive create role psp:unprivileged \ --verb=use \ --resource=podsecuritypolicy \ --resource-name=miztiik-secured # Create Role Binding kubectl -n miztiik-psp-restrictive create rolebinding sa-miztiik-dev:psp:unprivileged \ --role=psp:unprivileged \ --serviceaccount=miztiik-psp-restrictive:sa-miztiik-dev
-
Create Privileged Pod:
Save the following manifest as
busybox_privileged.yml
and deploy it,apiVersion: v1 kind: Pod metadata: name: busybox_privileged spec: hostNetwork: true containers: - name: busybox image: busybox command: ["sh", "-c", "sleep 1h"]
kubectl --as=system:serviceaccount:miztiik-psp-restrictive:sa-miztiik-dev \ -n miztiik-psp-restrictive \ apply -f busybox_privileged.yml
You should be seeing an error like this,
Error from server (Forbidden): error when creating "busybox_privileged.yml": pods "busybox_privileged" is forbidden: PodSecurityPolicy: unable to admit pod: [spec.securityContext.hostNetwork: Invalid value: true: Host network is not allowed to be used]
If you remove the line
hostNetwork: true
or change it to false and redeploy the manifest, for examplebusybox.yml
, This manifest will be deployed successfully.
-
-
Here we have demonstrated how to use Kubernetes Pod security Policy to scale kubernetes clusters, You can extend this solution to secure your cluster by adding more rules to your policy
-
If you want to destroy all the resources created by the stack, Execute the below command to delete the stack, or you can delete the stack from console as well
- Resources created during Deploying The Application
- Delete CloudWatch Lambda LogGroups
- Any other custom resources, you have created for this demo
# Delete from cdk cdk destroy # Follow any on-screen prompts # Delete the CF Stack, If you used cloudformation to deploy the stack. aws cloudformation delete-stack \ --stack-name "MiztiikAutomationStack" \ --region "${AWS_REGION}"
This is not an exhaustive list, please carry out other necessary steps as maybe applicable to your needs.
This repository aims to show how to use Pod Security Policy to new developers, Solution Architects & Ops Engineers in AWS. Based on that knowledge these Udemy course #1, course #2 helps you build complete architecture in AWS.
Thank you for your interest in contributing to our project. Whether it is a bug report, new feature, correction, or additional documentation or solutions, we greatly value feedback and contributions from our community. Start here
Buy me a coffee β.
- Kubernetes Docs: Pod Security Policies
- Kubernetes Docs: Security Context for a Pod
- AWS Docs: Amazon EKS default pod security policy
- Racher Docs: Pod Security Policies
- Racher Docs: Pod Security Policies - Dive Deeper
- Kubernetes Docs: Pod Security Policies Deprecation
- Kubernetes Docs: Pod Security Policies Deprecation
- AWS Blog: Amazon EKS Pod Security Policy
- AWS Blog: Amazon EKS Best Practices
Level: 300