CNCF Atlas

English

日本語

English

日本語

Appearance

Getting Started ​

Verified against the README at commit bd9c4c5 (near tag v2.2.3). The quickstart runs Higress standalone with Docker; the Kubernetes path uses Helm.

Prerequisites ​

  • Docker, for the standalone all-in-one quickstart.
  • For the Kubernetes path instead: a cluster and Helm 3.
  • Network access to Higress's own image registry (higress-registry.*.cr.aliyuncs.com). The README notes these images live in Higress's registry rather than Docker Hub, so Docker Hub rate limits do not apply. Pick the mirror closest to your region (cn-hangzhou, us-west-1, or ap-southeast-7).

Install ​

The fastest way to run Higress is the all-in-one Docker image, which bundles the control plane, the Envoy data plane, and the console. It writes its config to the mounted working directory, so no cluster is needed (README).

bash
mkdir higress; cd higress
docker run -d --rm --name higress-ai -v ${PWD}:/data \
  -p 8001:8001 -p 8080:8080 -p 8443:8443 \
  higress-registry.cn-hangzhou.cr.aliyuncs.com/higress/all-in-one:latest

The three ports are the console (8001), the gateway HTTP entry (8080), and the gateway HTTPS entry (8443) (README).

For a Kubernetes deployment, install with Helm and point global.hub at the mirror nearest you (README):

bash
helm install higress -n higress-system higress.io/higress \
  --set global.hub=higress-registry.us-west-1.cr.aliyuncs.com \
  --create-namespace

A first working setup ​

Using the standalone Docker install above:

  1. Start the all-in-one container with the docker run command from the Install section. It writes configuration files into the current directory.

  2. Open the console to confirm the control plane is up.

    bash
    open http://localhost:8001

  3. Send a request through the gateway HTTP entry. Before any route is configured the gateway answers on port 8080, which confirms the Envoy data plane is serving.

    bash
    curl -i http://localhost:8080/

From the console you can then add a route (a domain plus an upstream service) and see it take effect without restarting the gateway, which is the behavior the architecture page traces through xDS.

Verify it works ​

  • The container is running: docker ps shows higress-ai.
  • The console responds on http://localhost:8001.
  • The gateway data plane responds on port 8080 (an HTTP status back from curl, even a 404 before routes exist, means Envoy is up).

Where to go next ​

For Helm values, high availability, and production hardening on Kubernetes, follow the official Quick Start documentation. For the AI gateway, the ai-proxy plugin and its provider list live under plugins/wasm-go/extensions/ai-proxy/, and MCP server hosting is covered in the MCP quickstart. The Internals page maps the translation path if you want to read how Ingress becomes Envoy config.