Deploy to Kubernetes

To get started we recommend using the Golem Helm Chart (opens in a new tab). You will need to have Helm (opens in a new tab) and kubectl (opens in a new tab) installed locally and running kubernetes (opens in a new tab) cluster (for testing you can use Docker Desktop (opens in a new tab) or minikube (opens in a new tab)).

For quick start we recommend to download content of kube (opens in a new tab) directory from Golem repository (opens in a new tab) and run deploy.sh script:

./deploy.sh -n golem

which will deploy Golem with Redis and PostgreSQL to kubernetes namespace golem. Kubernetes Persistent Volumes (opens in a new tab) are used to store Golem's data. Nginx ingress controller (opens in a new tab) is used to expose Golem APIs (http with port 80). For testing, you may need to change golem-cli configuration to use proper host and port. Worker API Gateway endpoint is exposed under port 9006.

For production is preferable to use separate PostgreSQL cluster, Redis cluster and reliable Shared File System or S3 for Blob storage.

See Golem Helm Chart values.yaml (opens in a new tab) and readme (opens in a new tab) for more details.