Update on Jan 24th, 2023: this blog article is now on Medium.

At KubeCon, GitOpsCon, SigstoreCon and SecurityCon NA 2022, Secure Software Supply Chain (S3C) demonstrated that it is not anymore just a trend or a buzz. It’s getting more and more serious, we are seeing a lot of simplication about how to set up and leverage such technologies.

Don’t trust registries.

Sign everything: git commit, npm/rust/java/python packages, container image, Helm chart, Kubernetes manifests, etc.

Nearly every site runs HTTPS and is, by definition, more secure. This is the model that Sigstore wants to follow.

When I came back from KubeCon NA 2022, I added at the top of my TODO list to “play and learn more about Sigstore’s cosign in Kubernetes clusters. So here I am, like usual, sharing my step by step guide about how to accomplish this while sharing my thoughts and learnings. Hope you’ll like it and that you will learn something!

Note: while testing this feature, it was also the opportunity for me to open my first PRs in the sigstore/docs, sigstore/policy-controller, and sigstore/community repos to fix some frictions I faced: https://github.com/sigstore/docs/pull/63, https://github.com/sigstore/policy-controller/pull/520, and https://github.com/sigstore/community/issues/220.

This blog article consists on two main sections:

cosign with GKE and KMS flow

Define the common bash variables used throughout this blog article:

PROJECT_ID=FIXME-WITH-YOUR-EXISTING-PROJECT-ID
gcloud config set project ${PROJECT_ID}
REGION=northamerica-northeast1

Sign a container image with Cloud KMS and cosign

In this section you will:

  • Create a key in KMS
  • Create a Google Artifact Registry repository to store container images
  • Push a simple nginx container image in this repository
  • Install Sigstore’s cosign locally
  • Sign this remote private container image

Enable the KMS API in our current project:

gcloud services enable cloudkms.googleapis.com

Create a key in KMS:

KEY_RING=cosign
gcloud kms keyrings create ${KEY_RING} \
    --location ${REGION}
KEY_NAME=cosign
gcloud kms keys create ${KEY_NAME} \
    --keyring ${KEY_RING} \
    --location ${REGION} \
    --purpose asymmetric-signing \
    --default-algorithm ec-sign-p256-sha256

Enable the Artifact Registry API in our current project:

gcloud services enable artifactregistry.googleapis.com

Create a private Google Artifact Registry repository to store our container images:

REGISTRY_NAME=containers
gcloud artifacts repositories create ${REGISTRY_NAME} \
    --repository-format docker \
    --location ${REGION}

Push an nginx image in our own private Google Artifact Registry repository:

docker pull nginx
docker tag nginx ${REGION}-docker.pkg.dev/${PROJECT_ID}/${REGISTRY_NAME}/nginx
gcloud auth configure-docker ${REGION}-docker.pkg.dev
SHA=$(docker push ${REGION}-docker.pkg.dev/${PROJECT_ID}/${REGISTRY_NAME}/nginx | grep digest: | cut -f3 -d" ")

Note: we are grabbing the SHA of this remote container image in order to sign this container image later.

Install Sistore’s cosign locally:

COSIGN_VERSION=$(curl -s https://api.github.com/repos/sigstore/cosign/releases/latest | jq -r .tag_name)
curl -LO https://github.com/sigstore/cosign/releases/download/${COSIGN_VERSION}/cosign-linux-amd64
sudo mv cosign-linux-amd64 /usr/local/bin/cosign
chmod +x /usr/local/bin/cosign

Generage a key and sign this remote container image:

gcloud auth application-default login
cosign generate-key-pair \
    --kms gcpkms://projects/${PROJECT_ID}/locations/${REGION}/keyRings/${KEY_RING}/cryptoKeys/${KEY_NAME}
cosign sign \
    --key gcpkms://projects/${PROJECT_ID}/locations/${REGION}/keyRings/${KEY_RING}/cryptoKeys/${KEY_NAME} \
    ${REGION}-docker.pkg.dev/${PROJECT_ID}/${REGISTRY_NAME}/nginx@${SHA}

We could now see that our Google Artifact Registry repository has two entries, one for the actual container image and the other for the associate .sig signature:

gcloud artifacts docker tags list ${REGION}-docker.pkg.dev/${PROJECT_ID}/${REGISTRY_NAME}/nginx

Output similar to:

Listing items under project mabenoit-gatekeeper-oci, location us-east4, repository containers.
TAG                                                                          IMAGE                                                             DIGEST
latest                                                                       us-east4-docker.pkg.dev/mabenoit-gatekeeper-oci/containers/nginx  sha256:4c1c50d0ffc614f90b93b07d778028dc765548e823f676fb027f61d281ac380d
sha256-4c1c50d0ffc614f90b93b07d778028dc765548e823f676fb027f61d281ac380d.sig  us-east4-docker.pkg.dev/mabenoit-gatekeeper-oci/containers/nginx  sha256:f02d7fef0df5c264e34b995a4861590bbdd7001631f6e5f23250f34202359a56

Note: there is an ongoing discussion to support the reference types from the OCI spec in order to just have the container image where the signature could be attached on.

cosign verify \
    --key gcpkms://projects/${PROJECT_ID}/locations/${REGION}/keyRings/${KEY_RING}/cryptoKeys/${KEY_NAME} \
    ${REGION}-docker.pkg.dev/${PROJECT_ID}/${REGISTRY_NAME}/nginx@${SHA}

Output similar to:

Verification for us-east4-docker.pkg.dev/mabenoit-gatekeeper-oci/containers/nginx@sha256:4c1c50d0ffc614f90b93b07d778028dc765548e823f676fb027f61d281ac380d --
The following checks were performed on each of these signatures:
  - The cosign claims were validated
  - The signatures were verified against the specified public key

[{"critical":{"identity":{"docker-reference":"us-east4-docker.pkg.dev/mabenoit-gatekeeper-oci/containers/nginx"},"image":{"docker-manifest-digest":"sha256:4c1c50d0ffc614f90b93b07d778028dc765548e823f676fb027f61d281ac380d"},"type":"cosign container image signature"},"optional":null}]

Enforce that only signed container images are allowed in a GKE cluster with Sigstore’s policy-controller

In this section you will:

  • Create a dedicated least privilege Google Service Account for the GKE’s nodes
  • Create a GKE cluster
  • Install Sigstore’s policy-controller in this GKE cluster
  • Deploy a policy to only allow signed container images
  • Test this policy with both signed and unsigned container images

Define a least privilege Google Service Account (GSA) the GKE’s nodes (instead of using the default Compute Engine Service Account):

CLUSTER_NAME=cosign-test-cluster
GSA_NAME=${CLUSTER_NAME}-sa
GSA_ID=${GSA_NAME}@${PROJECT_ID}.iam.gserviceaccount.com
gcloud iam service-accounts create ${GSA_NAME} \
    --display-name ${GSA_NAME}
roles="roles/logging.logWriter roles/monitoring.metricWriter roles/monitoring.viewer roles/cloudkms.viewer roles/cloudkms.verifier"
for role in $roles; do gcloud projects add-iam-policy-binding ${PROJECT_ID} --member "serviceAccount:${GSA_ID}" --role $role; done
gcloud artifacts repositories add-iam-policy-binding ${REGISTRY_NAME} \
    --location ${REGION} \
    --member "serviceAccount:${GSA_ID}" \
    --role roles/artifactregistry.reader

Note: in addition to the roles for monitoring, we are also granting both cloudkms.viewer and cloudkms.verifier needed by Sigstore’s policy-controller.

Enable the GKE API in our current project:

gcloud services enable container.googleapis.com

Create a GKE cluster with this dedicated GSA:

gcloud container clusters create ${CLUSTER_NAME} \
    --service-account ${GSA_ID} \
    --region ${REGION} \
    --scopes "gke-default,https://www.googleapis.com/auth/cloudkms"

Note: we explicitly add the https://www.googleapis.com/auth/cloudkms scope needed by Sistore’s policy-controller. https://www.googleapis.com/auth/cloud-platform instead is fine too.

Install the Sigstore’s policy-controller Helm chart in this GKE cluster:

helm repo add sigstore https://sigstore.github.io/helm-charts
helm repo update
helm install policy-controller \
    -n cosign-system sigstore/policy-controller \
    --create-namespace

Deploy a policy only allowing signed container images from our private Google Artifact Registry repository:

cat << EOF | kubectl apply -f -
apiVersion: policy.sigstore.dev/v1alpha1
kind: ClusterImagePolicy
metadata:
  name: private-signed-images-cip
spec:
  images:
  - glob: "**"
  authorities:
  - key:
        kms: gcpkms://projects/${PROJECT_ID}/locations/${REGION}/keyRings/${KEY_RING}/cryptoKeys/${KEY_NAME}/cryptoKeyVersions/1
EOF

Enfore this policy for the test namespace:

kubectl create namespace test
kubectl label namespace test policy.sigstore.dev/include=true

Note: you need to apply this label on the namespaces you want this policy to be enforced in.

Test with an unsigned container image and see that it’s blocked:

kubectl create deployment nginx \
    --image=nginx \
    -n test

Output similar to:

error: failed to create deployment: admission webhook "policy.sigstore.dev" denied the request: validation failed: failed policy: private-signed-images-cip: spec.template.spec.containers[0].image
index.docker.io/library/nginx@sha256:b8f2383a95879e1ae064940d9a200f67a6c79e710ed82ac42263397367e7cc4e signature key validation failed for authority authority-0 for index.docker.io/library/nginx@sha256:b8f2383a95879e1ae064940d9a200f67a6c79e710ed82ac42263397367e7cc4e: no matching signatures:

Test with our signed container image and see that it’s allowed:

kubectl create deployment nginx \
    --image=${REGION}-docker.pkg.dev/${PROJECT_ID}/${REGISTRY_NAME}/nginx@${SHA} \
    -n test

Output similar to:

deployment.apps/nginx created

That’s it, congrats! We just enforced our GKE cluster to only allow our private and signed container images on specific namespaces! Wow!

Resources

Hope you enjoyed that one! Happy signing, happy sailing!