Monkeys
Start Here

Helm

Deploy Monkeys services to Kubernetes with Helm

Monkeys still uses the same product boundary in Kubernetes as it does locally. Core services stay together, while optional capability services are added only when needed.

Use Docker Compose for local development and smoke testing before creating Kubernetes values. Helm deployment should promote the same service boundaries to a cluster environment.

What to Deploy

AreaRepositories
Core productmonkeys-server, monkeys-studio, monkeys-kernel, monkeys-compute
Agentmonkeys-agent-server, monkeys-mcp-server, monkey-tools-agentkits
Billingmonkeys-billing-server
Datamonkeys-data-server, monkeys-conductor-worker
Webhookmonkeys-webhook-server
Workersmonkeys-conductor-worker, monkeys-training-worker, monkeys-comfyui-worker
Toolsmonkey-tools-third-party-api, monkey-tools-third-party-api-python, monkey-tools-sandbox, monkey-tools-agentkits

Reference Material

Deployment Checklist

  1. Prepare Kubernetes namespace, ingress, TLS, storage classes, and image pull credentials.
  2. Decide whether PostgreSQL, Redis, Elasticsearch, Conductor, MinIO, Prometheus, and Grafana are managed services or chart-managed dependencies.
  3. Build and publish service images for the repositories you deploy.
  4. Put runtime config and secrets in Kubernetes Secrets or your external secret store.
  5. Deploy core services first, then add optional capability services.
  6. Verify health, metrics, workflow execution, and Studio API connectivity.

Notes

  • The docs repo intentionally does not keep production secrets.
  • Runtime config such as config.yaml stays in the owning service repository or the deployment secret store.
  • Use the service README and deployment charts together when deploying a specific capability.

On this page