Skip to main content

Prerequisites

  • Kubernetes 1.25+
  • Helm 3.8+

Install

This deploys a single-instance Proxy Hopper with the in-memory backend using the default config.

Supplying your config

Inline config (simplest)

Pass your config.yaml content as a Helm value:
Or in a values.yaml file:

Existing ConfigMap or Secret

If you manage config separately (via an operator, external-secrets, or sealed-secrets), reference it:
When auth is enabled in your config (API keys, JWT, OIDC), use existingSecret rather than config.inline in a values file. Credentials in plain values files can end up in version control or Helm release history.Create the secret with:

Redis backend

Enable the bundled Redis subchart for pool state persistence and multi-instance HA:
The chart automatically selects the -redis Docker image and wires the Redis URL when backend.type: redis is set. To use an external Redis instance instead of the bundled subchart:

Token Server

If you’re using managed auth, the chart can deploy your token server image alongside Proxy Hopper — a Deployment plus a ClusterIP Service, with a predictable in-cluster URL derived automatically:
Wire it into Proxy Hopper’s own config by referencing the chart’s URL helper — this keeps the URL correct regardless of the Helm release name:
If you’re using config.existingConfigMap / config.existingSecret instead of config.inline, the template helper isn’t available — hard-code the in-cluster URL instead:
Your token server image only needs to implement POST /token and GET /health — see Building a Token Server for the contract and a Python library that implements it for you.

Ingress

Upgrading

Uninstall