Consul API Gateway 0.3.0

Release Highlights

Improved High Availability and Scalability via Multiple Instances per Gateway: This release introduces support for deploying multiple instances of a logical gateway (i.e. a K8s gateway config object). You can use an external load balancing service to distribute incoming traffic across the instances of a gateway. This provides higher availability of gateway services and also allows a single logical gateway to handle a volume of network traffic that is much greater than what a single instance of a gateway can.

Administrators can set the number of instances independently for each logical gateway. Administrators can also set, in the gatewayClassConfig, the minimum, maximum and default number of gateway instances. These are applied per gateway to all gateways in that gatewayClass. The minimum and maximum settings will override settings configured on any given gateway.

Note: There is a hard limit of eight (8) instances per gateway and it can not be overridden. This hard limit will probably be increased in the future.
Cross Namespace Reference Policies for Certificates: Reference policies is a security feature introduced in the 0.2 release. In this release, Reference Policies are now required for TLS certificates when a listener has a certificateRef that is in a different namespace.

Consul 1.11.2+
HashiCorp Consul Helm chart 0.45.0+
Kubernetes 1.21+
- Kubernetes 1.24 is not supported at this time.
Kubectl 1.21+
Envoy proxy support is determined by the Consul version deployed. Refer to Envoy Integration for details.

Supported version of the Gateway API spec: v1alpha2(v0.4.1)

For detailed information on upgrading, please refer to the upgrade details page

The changelogs for this major release version and any maintenance versions are listed below.

Note: The following link will take you to the changelogs on the GitHub website.