Skip to main content

AI/Run CodeMie Components Deployment

Overview

This section describes the process of deploying the main AI/Run CodeMie components to the AWS EKS cluster.

info

For infrastructure deployment details, refer to the Infrastructure Deployment section.

Prerequisites

  1. Obtain kubeconfig for created cluster:

    aws eks update-kubeconfig --region <REGION> --name <PLATFORM_NAME>
  2. Ensure EKS cluster has installed:

    • Nginx Ingress Controller
    • AWS gp3 storage class
info

If your EKS cluster does not have these components, the detailed instructions are provided in the Manual Components Installation section.

  1. Clone the codemie-helm-charts repository

  2. Set up pull secrets for container registry access. For detailed instructions, see Container Registry Access.

Application Stack Overview

AI/Run CodeMie Application Stack

The AI/Run CodeMie deployment includes:

Core Components

info

AI/Run CodeMie latest releases for core components versions can be found by executing following script in the codemie-helm-charts repository for each component.

bash get-codemie-latest-release-version.sh
bash get-codemie-latest-release-version.sh -c ./path/to/key.json

Make sure you logged in with key.json shared with you.

info

Versions for Docker containers and Helm releases are matching

Component nameImagesDescription
AI/Run CodeMie APIeurope-west3-docker.pkg.dev/or2-msq-epmd-edp-anthos-t1iylu/prod/codemie:x.y.zThe backend service of the AI/Run CodeMie application responsible for business logic, data processing, and API operations
AI/Run CodeMie UIeurope-west3-docker.pkg.dev/or2-msq-epmd-edp-anthos-t1iylu/prod/codemie-ui:x.y.zThe frontend service of the AI/Run CodeMie application that provides the user interface for interacting with the system
AI/Run CodeMie Nats Auth Callouteurope-west3-docker.pkg.dev/or2-msq-epmd-edp-anthos-t1iylu/prod/codemie-nats-auth-callout:x.y.zAuthorization component of AI/Run CodeMie Plugin Engine that handles authentication and authorization for the NATS messaging system
AI/Run CodeMie MCP Connecteurope-west3-docker.pkg.dev/or2-msq-epmd-edp-anthos-t1iylu/prod/codemie-mcp-connect-service:x.y.zA lightweight bridge tool that enables cloud-based AI services to communicate with local Model Context Protocol (MCP) servers via protocol translation while maintaining security and flexibility
AI/Run Mermaid Servereurope-west3-docker.pkg.dev/or2-msq-epmd-edp-anthos-t1iylu/prod/mermaid-server:x.y.zImplementation of open-source service that generates image URLs for diagrams based on the provided Mermaid code for workflow visualization

Required Third-Party Components

Component nameImagesDescription
Ingress Nginx Controllerregistry.k8s.io/ingress-nginx/controller:x.y.zHandles external traffic routing to services within the Kubernetes cluster. The AI/Run CodeMie application uses oauth2-proxy, which relies on the Ingress Nginx Controller for proper routing and access control
Storage ClassProvides persistent storage capabilities
Elasticsearchdocker.elastic.co/elasticsearch/elasticsearch:x.y.zDatabase component that stores all AI/Run CodeMie data, including datasources, projects, and other application information
Kibanadocker.elastic.co/kibana/kibana:x.y.zWeb-based analytics and visualization platform that provides visualization of the data stored in Elasticsearch. Allows monitoring and analyzing AI/Run CodeMie data
Postgres-operatorregistry.developers.crunchydata.com/crunchydata/postgres-operator:x.y.zManages PostgreSQL database instances required by other components in the stack. Handles database lifecycle operations
Keycloak-operatorepamedp/keycloak-operator:x.y.zManages Keycloak identity and access management instance and it's configuration
Keycloakdocker.io/busybox:x.y.z, quay.io/keycloak/keycloak:x.y.z, registry.developers.crunchydata.com/crunchydata/crunchy-postgres:x.y.zIdentity and access management solution that provides authentication and authorization capabilities for integration with oauth2-proxy component
Oauth2-Proxyquay.io/oauth2-proxy/oauth2-proxy:x.y.zAuthentication middleware that provides secure authentication for the AI/Run CodeMie application by integrating with Keycloak or any other IdP
NATSnats:x.y.z, natsio/nats-server-config-reloader:x.y.zMessage broker that serves as a crucial component of the AI/Run CodeMie Plugin Engine, facilitating communication between services
LLM ProxyOptional proxy component that balances requests to Azure OpenAI language models (LLMs), providing high availability and load distribution
Fluent Bitcr.fluentbit.io/fluent/fluent-bit:x.y.zFluent Bit enables logs and metrics collection from AI/Run CodeMie enabling the Agents observability

Deployment Methods

Choose your preferred deployment method:

Finalizing Installation

After component deployment, you should have access to:

ComponentURL Pattern
AI/Run CodeMie UIhttps://codemie.<your-domain>
AI/Run CodeMie APIhttps://codemie.<your-domain>/code-assistant-api/v1
Keycloak UIhttps://keycloak.<your-domain>/auth/admin
Kibanahttps://kibana.<your-domain>
info

Some components may be missing depending on your setup configuration or use http protocol in private clusters.

Next Steps

After successful components deployment, proceed to Post-Installation Configuration to complete required setup steps.