Skip to main content

Setup guide

Kubernetes Cluster Installation​

Rig shines when it gets to run in a real Kubernetes environment. It lowers the barrier of entry for engineers as they won't have to learn the complex APIs of Kubernetes.

Rig consists of two components which together gives the full experience of Rig.dev.

  • rig-operator: our open source components which provide higher level APIs in kubernetes, which encapsulate multiple low level resources in a single CRD.
  • rig-platform: our paid product which further builds on the primitives provided by rig-operator.

Each of these components have their own requirements and offer options for configuration to make it fit into your kubernetes cluster.

Requirements: rig-operator​

The rig-operator uses the kubernetes APIs and a few additional APIs to provide some very easy-to-use APIs for spinning up workloads.

cert-manager​

The rig-operator helm chart uses cert-manager to create a certificate for the CRD webhooks. These are self-signed and doesn't require any special configuration of cert-manager.

If you want to enable ingress support, then you will need to further configure cert-manager and rig-operator to be able to generate certificates for workloads. See the section about ingress for more information about this.

metrics-server​

The rig-operator creates HorizontalPodAutoscaler resources, configured to use the metrics APIs provided by metrics-server.

In some environments (eg. GKE) this is installed in the kubernetes cluster when you create it. To check wether the metrics APIs are available run

kubectl api-resources | grep metrics

This should show an output similar to the below if the APIs are available.

nodes          metrics.k8s.io/v1beta1       false        NodeMetrics
pods           metrics.k8s.io/v1beta1       true         PodMetrics

If you dont see this output you can install the metrics-server using the official instructions.

Requirements: rig-platform​

The rig-platform needs a database for holding its data and the metrics APIs for showing graphs in the dashboard, and then it needs the rig-operator and the APIs it provides.

If you installed rig-operator and its requirements, then you already have the metrics APIs available.

Database​

Support is available for connecting to an existing PostgreSQL or MongoDB database. Alternatively, a PostgreSQL database can be deployed for you as part of the installation (not recommended for production environments).

Installation​

We will install the Rig Platform using Helm. This also installs the Rig Operator, which enables support for Capsules within the cluster.

helm upgrade --install rig-operator rig-operator \
  --repo https://charts.rig.dev \
  --version 0.1.7 \
  --namespace rig-system \
  --create-namespace
helm upgrade --install rig-platform rig-platform \
  --repo https://charts.rig.dev \
  --version 0.2.7 \
  --namespace rig-system \
  --create-namespace \
  --set postgres.enabled=true  # Remove to not start a new PostgreSQL database.
Note

See the Database article for more information about how to connect to your own database, and other configuration options.

After installing the Rig Platform you'll need to run the following command to setup an admin user

kubectl exec -it --namespace rig-system deploy/rig-platform \
  -- rig-admin init

Configuration​

The Rig Operator and Platform both have a values.yaml file which specify how to configure their respective Helm charts. The values files and their defaults can be found here

To overwrite a default value when deploying, use --set FIELD=VALUE flags when running helm.

Connect to the Rig Platform remotely​

After setting up Rig, users who want to connect to it remotely can run the following command using their local Rig CLI

rig config init

which will guide the user trough connecting to Rig.