Open Policy Agent Gatekeeper
open-policy-agent.github.io
- Free plan
- Open source
- Linux
- API
- Self-hosted
What Open Policy Agent Gatekeeper is
Open Policy Agent Gatekeeper is a free Kubernetes policy controller that applies rules defined through custom resource definitions and evaluated by Open Policy Agent. It operates as a validating and mutating webhook: validation constraints can deny, warn, run in dry-run mode, or report violations, while mutation policies can alter matching resources before validation checks the final object. Its audit function also periodically checks existing cluster resources against constraints to surface earlier misconfigurations. The project provides an extensible policy library with parameterized samples that users can adapt, as well as provider-based connections to external sources such as image registries and directory services. Named integrations or backed services include Azure Policy for Kubernetes, Google Kubernetes Engine Policy Controller, Rancher, AWS EKS Blueprints, and Alibaba Cloud. Gatekeeper is self-hosted for Linux, with deployment options including Helm and several image-build approaches. Installation requires cluster-admin permissions and Kubernetes resources introduced in version 1.16. Audit status reports show up to 20 violations by default, though that limit can be configured; the documentation notes object-size and memory constraints.
Who it suits
Gatekeeper suits Kubernetes administrators and platform teams that need policy enforcement and auditing for cluster resources. It is aimed at users able to install and operate self-hosted software with cluster-admin permissions.
Good to know, and what to watch
- Validation can deny, warn, or run in dry-run mode.
- Mutation can change resources before validation.
- Audit checks existing resources for misconfigurations.
- Policy examples can be extended and customized.
- Deployment options include Helm and container images.
- Installation requires cluster-admin permissions.
- Requires Kubernetes resources introduced in version 1.16.
- Audit reports show 20 violations by default.
Plans
Every plan and priceGatekeeper
Open-source Gatekeeper
Free plan
The verdict
Gatekeeper combines admission-time policy enforcement with recurring checks of existing cluster resources. Teams should account for its installation permissions and the default cap on reported audit violations.
What the maker publishes
- API and extensibility
- The documentation exposes an API reference and describes Kubernetes CRDs, runtime flags, mutators, external-data providers, and generated ValidatingAdmissionPolicies as extension points. source
- Audit
- Its audit function periodically evaluates existing cluster resources against constraints to detect pre-existing misconfigurations. source
- Audit
- The audit operation periodically scans existing cluster objects and writes discovered violations to constraint status. source
- Audit limit
- Constraint status reports at most 20 individual violations by default, configurable with a runtime flag; the documentation notes Kubernetes object-size and memory limits. source
- Build security
- Each image has build attestations, an SBOM, and SLSA provenance; attestations are available for releases starting with v3.12.0. source
- CLI
- The gator CLI is documented as a beta feature beginning with Gatekeeper version 3.11. source
- CLI
- The project documents the gator command-line interface for validating workload resources and policies. source
- Cloud integrations
- The project site names Azure Policy for Kubernetes, Google Kubernetes Engine Policy Controller, Rancher, AWS EKS Blueprints, and Alibaba Cloud as Gatekeeper integrations or backed services. source
- Cloud integrations
- The official site lists Azure Policy for Kubernetes, Google Kubernetes Engine Policy Controller, Rancher, AWS EKS Blueprints, and Alibaba Cloud Kubernetes as integrations or backed services. source
- Community support
- The project site links to a Gatekeeper Slack channel and community meetings. source
- Constraints
- Constraint resources instantiate templates with matching rules and parameters for selected Kubernetes objects. source
- Constraint templates
- ConstraintTemplate resources contain Rego logic and schemas that define reusable policy types. source
- Documentation depth
- The documentation includes installation, examples, audit, mutation, external data, debugging, metrics, performance, architecture, runtime flags, API reference, and contribution material. source
- External data
- Gatekeeper supports provider-based connections to external data sources, including image registries and directory services. source
- External data
- Gatekeeper supports external-data providers that let admission policies query data outside the Kubernetes object. source
- Independent limitation: scope
- HackingNote characterizes Gatekeeper as Kubernetes-only, unlike general-purpose OPA deployments. source
- Independent summary
- Hysen Labs describes Gatekeeper as a self-hosted Kubernetes service written mainly in Go with Apache-2.0 licensing. source
- Independent use cases
- A Kubernetes project article gives examples including approved image repositories, globally unique ingress hostnames, pod resource limits, and namespace ownership labels. source
- Installation
- Installation supports prebuilt release images, development images, locally built images, and Helm-based deployment. source
- Install options
- The documentation describes deployment using a prebuilt image, a development image, a locally built image, or Helm. source
- Kubernetes requirement
- Gatekeeper requires Kubernetes resources introduced in version 1.16 and requires cluster-admin permissions for installation. source
- Licence
- The Gatekeeper repository is licensed under Apache-2.0. source
- Limit: complexity
- The 2025 survey reports that some users find Gatekeeper overcomplicated for simple policies and find constraint-template debugging difficult. source
- Limit: learning curve
- A production user reports that implementation can be difficult for unprepared organizations because Gatekeeper adds components and requires learning Rego. source
Checked against the maker's pages on 15 Oct 2026.