Skip to main content

Services

This section introduces the Datalayer services you deploy on top of your Kubernetes cluster.

Use it as your service map before running Plane deployment commands.

Services are grouped in three layers:

  1. System services: cluster foundations such as ingress, certificates, storage and observability.
  2. Core services: IAM, operator, runtimes, spacer, contents, library, AI agents, Jupyter MCP Server, functions, scheduler, spider, manager, status.
  3. Optional and addons: extra capabilities you enable depending on your use case.
  1. System Layer First:
    • Cert Manager
    • Ingress
    • Solr
    • OTEL
    • Durable — workflows that outlive the process that started them; deploy it after the PostgreSQL operator and before the gateway, on the runtimes plane
    • Node Mounts — the privileged node DaemonSet holding the two below; off by default
      • Node Mount Gateway — home folders, datasets, NFS, Git and buckets into Pods that are already running
      • Local CSI driver — local.csi.datalayer.io, a folder of a user's own computer inside a sandbox
  2. Core Platform Layer:
  3. Optional Services and Addons:

Minimum Service Set​

For a base deployment, start with:

  1. Cert Manager
  2. Ingress
  3. Solr
  4. Datalayer Operator
  5. Datalayer IAM
  6. Datalayer Runtimes
  7. Datalayer Contents
  8. Datalayer AI Agents

Each service page below documents purpose, deployment details, and operational checks.

Terraform-First Workflow (AWS)​

Terraform-First Workflow (AWS)

Before deploying services with Plane, provision the infrastructure and Kubeadm cluster:

cd terraform
terraform init
terraform apply
./generated/clouder-kubeadm-setup.sh
export KUBECONFIG=~/.clouder/kubeadm/$(terraform output -raw cluster_name)/kubeconfig
kubectl get nodes -o wide

Then deploy all services in one pass:

./generated/plane-deploy-all-services.sh