Skip to main content

Gateway installation overview

Gateway resources expose and manage AI traffic to the data plane. This page describes the tare install-gateway command and its workflow; Agent Router–specific runbooks follow.

Synopsis​

tare install-gateway <identity-file> [flags]

Description​

The tare install-gateway command deploys the gateway Helm chart required to expose Agent Router serve traffic. It manages the lifecycle of the gateway resources, from prerequisite validation to final DNS configuration guidance.

High-level workflow​

  1. Identity loading: Validates and loads the provided <identity-file>.
  2. Configuration merging: Loads the optional gcp-gateway.json via --config and resolves values across flags, config, and identity defaults.
  3. Prerequisite handling: Runs optional prerequisite workflows with --dry-run-prereqs or --apply-prereqs.
  4. Preflight validation: Verifies environment readiness.
  5. Installation: Deploys the Agent Router gateway chart through the embedded Helm engine.
  6. Finalization: Optionally waits for the Gateway address (--wait) and outputs the required DNS A-record actions.

Provider support​

ProviderStatus
Google Cloud Platform (GCP)Fully supported
Azure Kubernetes Service (AKS)Fully supported
Amazon Web Services (AWS)Fully supported
note

The AWS path is ingress-only and takes no config file. Instead of deploying the gateway chart, it exposes the data plane's egress as a Network Load Balancer by patching the egress EnvoyProxy, so no AWS Load Balancer Controller and no aws CLI are required.

Where to go next

Gateway installation on GCP

Install the Agent Router gateway in a Google Cloud project using the tare CLI.

Gateway installation on AKS

Install the Agent Router gateway on Azure Kubernetes Service using the tare CLI.

Gateway installation on EKS

Expose the data plane on Amazon EKS as a Network Load Balancer using the tare CLI.