Introduction

Galasa provides Helm charts to install various components, the main one being a Galasa Ecosystem.

Prerequisites

Note: The Galasa Ecosystem chart only supports x86-64 at the moment. It cannot be installed on ARM64-based systems.

Helm must be installed to use the charts. Please refer to Helm's documentation to get started.

Galasa Ecosystem chart

Minikube

It is highly discouraged to use minikube for production purposes since it only provides a single Kubernetes node and will not scale well in demanding situations. Only use minikube for development and testing purposes.

If you would like to install the chart into minikube, ensure you have minikube installed and that it is running with minikube status. If minikube is not running, start it by running minikube start.

Once minikube is running, follow the instructions in the sections below to install the Galasa Ecosystem chart.

RBAC

If RBAC is active on your Kubernetes cluster, you will need to get your Kubernetes administrator to replace the placeholder username in the rbac-admin.yaml file with a username corresponding to a user with access to your cluster to assign them the galasa-admin role. This role allows assigned users to run the helm install/upgrade/delete commands to interact with the helm chart.

If multiple users require admin privileges, multiple groups, users, or ServiceAccounts can be assigned the galasa-admin role by extending the subjects list (see Using RBAC Authorization for more information).

For chart versions following 0.23.0, the other RBAC file (rbac.yaml) is applied automatically when installing the ecosystem chart. It creates a Galasa service account if one does not already exist so the API, Engine Controller, Metrics, and Resource Monitor can coordinate, while allowing the Engine Controller to create and manage engine pods.

For chart version 0.23.0 and prior, you will need to apply rbac.yaml manually. You can do this by running the following command:

kubectl apply -f https://raw.githubusercontent.com/galasa-dev/helm/ecosystem-0.23.0/charts/ecosystem/rbac.yaml

Installation

First, add the galasa repository as follows:

helm repo add galasa https://galasa-dev.github.io/helm

If you have already added this repository earlier, run helm repo update to retrieve the latest versions of the packages. You can then run helm search repo galasa to see the available charts.

Note: The Galasa Ecosystem Helm chart will deploy three persistent volumes. If you need to provide a Kubernetes storage class for these PVs, update the storageClass value in your values.yaml file with the name of a valid StorageClass on your cluster.

Download the values.yaml file and within it:

1. Set the galasaVersion value to a version of galasa you want to run (see releases for released versions). You should not use latest to ensure each pod in the Ecosystem is running at the same level.
2. Set the externalHostname value to the hostname that will be used to access Galasa services.
   • If you are deploying to minikube, the cluster's IP address can be retrieved by running minikube ip.

Once you have updated the galasaVersion and externalHostname values, continue following the instructions below to set up Ingress and Dex for your ecosystem.

Configuring Ingress

The ecosystem chart uses Ingress to reach services running within a Kubernetes cluster. See the Kubernetes documentation to learn more about Ingress.

Assuming your Ingress controller has been set up on your Kubernetes cluster, update the values under the ingress section within your values.yaml file as follows to configure the use of Ingress in your ecosystem:

1. Replace the ingressClassName value with the name of the IngressClass that is configured in your cluster. By default, nginx is used.

2. If you are using HTTPS, add a tls configuration within the ingress section, specifying the hosts list and a secretName value corresponding to the name of the Kubernetes Secret that contains your TLS private key and certificate. See the Kubernetes documentation for information for how to set up TLS.

Configuring Dex

As of Galasa version 0.32.0, Dex is used to authenticate users attempting to interact with a Galasa Ecosystem.

To configure Dex in your ecosystem, update your values.yaml file according to the following steps:

1. Replace the hostname in your issuer value with the same hostname given in externalHostname and set the URI scheme to either http or https. For example:

   issuer: http://<your-external-hostname>/dex

2. If desired, update the expiry section to configure the expiry of JSON Web Tokens (JWTs) and refresh tokens issued by Dex. By default, JWTs expire 24 hours after being issued and refresh tokens remain valid unless they have not been used for one year. See the Dex's documentation on ID tokens for information and available expiry settings.

Next, you will need to configure Dex to authenticate via a connector to authenticate with an upstream identity provider like GitHub, Microsoft, or an LDAP server. For a full list of supported connectors, refer to the Dex documentation. In this guide, we will configure Dex to authenticate through GitHub:

1. Register an OAuth application in GitHub, ensuring the application's callback URL is set to your Dex issuer value followed by /callback. For example, if your issuer value is https://prod-ecosystem.galasa.dev/dex, then your callback URL would be https://prod-ecosystem.galasa.dev/dex/callback.

2. Add a GitHub connector to your Dex configuration, providing the name of your GitHub organisation and any teams that you require users to be part of to be able to use your ecosystem as follows:

   dex:
     config:
       # Other Dex configuration values...
   
       connectors:
       - type: github
         id: github
         name: GitHub
         config:
           clientID: $GITHUB_CLIENT_ID
           clientSecret: $GITHUB_CLIENT_SECRET
           redirectURI: <your-dex-issuer-url>/callback
           orgs:
           - name: my-org
             teams:
             - my-team

   where $GITHUB_CLIENT_ID and $GITHUB_CLIENT_SECRET correspond to the registered OAuth application's client ID and secret. Also ensure that the redirectURI value is the same value that you provided when setting up your GitHub OAuth application in step 1.

   If you would like to pull the client ID and secret values of your OAuth application from a Kubernetes Secret, create a Secret by running the following kubectl command, ensuring the Secret's keys match those given in the GitHub connector's clientID and clientSecret values without the leading $ (i.e. GITHUB_CLIENT_ID and GITHUB_CLIENT_SECRET in the following example):

   kubectl create secret generic my-github-oauth-app-credentials \
   --from-literal=GITHUB_CLIENT_ID="myclientid" \
   --from-literal=GITHUB_CLIENT_SECRET="myclientsecret"

   Once your Kubernetes Secret has been created, you can supply the name of the Secret using the envFrom value in your values.yaml file to mount the Secret as follows:

   dex:
     envFrom:
       - secretRef:
         name: my-github-oauth-app-credentials
   
     config:
       # Other Dex configuration values...
   
       connectors:
       - type: github
         id: github
         name: GitHub
         config:
           clientID: $GITHUB_CLIENT_ID
           clientSecret: $GITHUB_CLIENT_SECRET
           redirectURI: <your-dex-issuer-url>/callback
           orgs:
           - name: my-org
             teams:
             - my-team

By default, the Galasa Ecosystem Helm chart will create a Kubernetes Secret containing configuration details for Dex. If you would like to apply your own Dex configuration as a Secret, your Dex configuration must be provided in a config.yaml key within the Secret and the value of the config.yaml key must be a valid Dex configuration.

For more information on configuring Dex, refer to the Dex documentation.

Having configured your values.yaml file, use the following command to install the Galasa Ecosystem Helm chart:

helm install -f /path/to/values.yaml <release-name> galasa/ecosystem --wait

where /path/to/values.yaml is the path to the values.yaml file that you downloaded, and <release-name> is the name that you want to give the ecosystem.

The --wait flag ensures the chart installation has completed before marking it as "Deployed". During the installation, the API pod waits for the etcd and RAS pods to initialise while the engine-controller, metrics, and resource-monitor pods wait for the API pod to initialise.

You can view the status of the deployed pods at any time by running kubectl get pods in another terminal. The results should look similar to the following:

NAME                                      READY   STATUS     RESTARTS      AGE
test-api-7945f959dd-v8tbs                 1/1     Running    0             65s
test-dex-5dc7fcb55f-lqv6s                 1/1     Running    0             65s
test-engine-controller-56fb476f45-msj4x   0/1     Init:0/1   0             65s
test-etcd-0                               1/1     Running    0             65s
test-metrics-5fd9f687b6-rwcww             0/1     Init:0/1   0             65s
test-ras-0                                1/1     Running    0             65s
test-resource-monitor-778c647995-x75z9    0/1     Init:0/1   0             65s
test-webui-6c896974d8-2k2tk               1/1     Running    0             65s

Verifying your Galasa Ecosystem Installation

After the helm install command ends with a successful deployment message, you can run the following command to ensure the Ecosystem can be accessed externally to Kubernetes and a simple test engine can be run:

helm test <release-name>

where <release-name> is the name that you gave the ecosystem during installation.

Once the helm test command ends and displays a success message, the Ecosystem has been set up correctly and is ready to be used.

Accessing services

Using Ingress, the URL of the Ecosystem bootstrap will be your external hostname followed by /api/bootstrap.

For example, if the external hostname you provided was example.com and you have provided values for using TLS, the bootstrap URL would be https://example.com/api/bootstrap. This is the URL that you would enter into a galasactl command's --bootstrap option to interact with your ecosystem.

Upgrading the Galasa Ecosystem

If you want to upgrade the Galasa Ecosystem to use a newer version of Galasa, for example, then you can use the following commands:

helm repo update
helm upgrade <release-name> galasa/ecosystem --reuse-values --set galasaVersion=0.33.0 --wait

Development

To install the latest development version of the Galasa Ecosystem chart, clone this repository and update the following values in your values.yaml file:

1. Set the galasaVersion value to main
2. Set the galasaRegistry value to harbor.galasa.dev/galasadev
3. Set the externalHostname value to the hostname that will be used to access Galasa services.
   • When deploying to minikube on Linux/macOS, add an entry to your /etc/hosts file like the one shown below, ensuring the IP address matches the output of minikube ip.

     192.168.49.2 example.com

Follow the installation instructions above to update the rest of your values.yaml file, including values to configure Ingress and Dex.

If you are deploying to minikube, ensure the NGINX Ingress controller is enabled by running:

minikube addons enable ingress

Once you have updated your values.yaml file, run the following command, providing the path to the ecosystem directory in this repository (e.g. ~/helm/charts/ecosystem).

helm install <release-name> /path/to/helm/charts/ecosystem --wait

Once the helm install command ends with a successful deployment message, you can follow the installation instructions above to test the deployed ecosystem using helm test and determine the bootstrap URL.