Chainguard Container for kserve

Chainguard Containers are regularly-updated, secure-by-default container images.

Download this Container Image

For those with access, this container image is available on cgr.dev:

docker pull cgr.dev/ORGANIZATION/kserve:latest

Be sure to replace the ORGANIZATION placeholder with the name used for your organization's private repository within the Chainguard Registry.

Usage

KServe enables the scalable deployment of machine learning models in Kubernetes. It provides mechanisms for model versioning, monitoring, and autoscaling, and is built on top of Kubernetes' extensibility features, such as custom resource definitions (CRDs) and operators.

Components

This deployment includes the following critical KServe components:

• Agent: Responsible for dynamically pulling models from storage (e.g., Google Cloud Storage or S3) and loading them into the model server.
• KServe Controller: Manages the lifecycle of ML models and their associated services in Kubernetes.
• ModelMesh Controller: Manages model serving infrastructure, including dynamic model loading and inference.
• ModelMesh Runtime Adapter: Communicates between ModelMesh and models.
• REST Proxy: Acts as a proxy layer for model serving, supporting multiple model types.
• Router: Manages routing inference requests to the appropriate model version or model instance. It handles traffic distribution, enabling versioned model serving and A/B testing.
• Storage Initializer: Downloads models from storage systems and prepares them for inference.

Installation

You can install KServe components using Helm charts. Below are instructions for deploying KServe with Chainguard images.

1. Install KServe CRDs

KServe relies on Custom Resource Definitions (CRDs) to manage model serving resources.

helm install kserve-crd oci://ghcr.io/kserve/charts/kserve-crd --version v0.13.1

2. Deploy KServe Components with Custom Images

To deploy KServe with Chainguard images, use the following Helm commands:

helm install kserve oci://ghcr.io/kserve/charts/kserve --version v0.13.1 \
  --set kserve.agent.image.repository=cgr.dev/chainguard/kserve-agent \
  --set kserve.agent.image.tag=latest
  --set kserve.router.image.repository=cgr.dev/chainguard/kserve-router \
  --set kserve.router.image.tag=latest
  --set kserve.controller.deploymentMode="RawDeployment" \
  --set kserve.controller.image.repository=cgr.dev/chainguard/kserve-controller \
  --set kserve.controller.image.tag=latest \
  --set kserve.modelmesh.enabled=true \
  --set kserve.modelmesh.controller.image.repository=cgr.dev/chainguard/kserve-modelmesh-controller \
  --set kserve.modelmesh.controller.image.tag=latest \
  --set kserve.modelmesh.config.restProxyImage=cgr.dev/chainguard/kserve-rest-proxy \
  --set kserve.modelmesh.config.restProxyImageTag=latest \
  --set kserve.storage.image.repository=cgr.dev/chainguard/kserve-storage-initializer \
  --set kserve.storage.image.tag=latest
  --set kserve.modelmesh.config.modelmeshImage=cgr.dev/chainguard/kserve-modelmesh \
  --set kserve.modelmesh.config.modelmeshImageTag=latest \
  --set kserve.modelmesh.config.modelmeshRuntimeAdapterImage=cgr.dev/chainguard/kserve-modelmesh-runtime-adapter \
  --set kserve.modelmesh.config.modelmeshRuntimeAdapterImageTag=latest \

For detailed installation configurations, refer to the KServe Helm chart values.yaml.

Inference Testing

Inference tests are critical to validating KServe's model serving capabilities. After deploying KServe, you can run inference tests to confirm that the deployed models are working as expected. For more info

1. Apply a Sample Inference Service

The following example creates an inference service that uses the iris-sklearn model:

kubectl apply -f /tests/inference-service.yaml

This service will download the model using the storage-initializer, and KServe will expose the model for inference requests.

2. Test Inference Using Port Forwarding

Once the inference service is running, you can send inference requests to validate the model's predictions:

kubectl port-forward -n kserve-test svc/iris-sklearn-predictor 8080:80 &

Use the following command to send a test prediction request:

curl -s -X POST http://localhost:8080/v1/models/iris-sklearn:predict \
  -H "Content-Type: application/json" \
  -d '{"instances": [[5.1, 3.5, 1.4, 0.2]]}'

You should receive a response like this:

{"predictions":[0]}

This confirms that the inference service is functioning correctly.

What are Chainguard Containers?

Chainguard Containers are minimal container images that are secure by default.

In many cases, the Chainguard Containers tagged as :latest contain only an open-source application and its runtime dependencies. These minimal container images typically do not contain a shell or package manager. Chainguard Containers are built with Wolfi, our Linux undistro designed to produce container images that meet the requirements of a more secure software supply chain.

The main features of Chainguard Containers include:

• Minimal design, without unnecessary software bloat
• Daily builds to ensure container images are up-to-date with available security patches
• High quality build-time SBOMs attesting to the provenance of all artifacts within the image
• Verifiable signatures provided by Sigstore
• Reproducible builds with Cosign and apko (read more about reproducibility)

For cases where you need container images with shells and package managers to build or debug, most Chainguard Containers come paired with a -dev variant.

Although the -dev container image variants have similar security features as their more minimal versions, they feature additional software that is typically not necessary in production environments. We recommend using multi-stage builds to leverage the -dev variants, copying application artifacts into a final minimal container that offers a reduced attack surface that won’t allow package installations or logins.

Learn More

To better understand how to work with Chainguard Containers, please visit Chainguard Academy and Chainguard Courses.

In addition to Containers, Chainguard offers VMs and Libraries. Contact Chainguard to access additional products.

Trademarks

This software listing is packaged by Chainguard. The trademarks set forth in this offering are owned by their respective companies, and use of them does not imply any affiliation, sponsorship, or endorsement by such companies.