# Helm Charts

DataTorch uses Helm 3 when deploying to a Kubernetes cluster. This provided a simple package manager to handle all of the required services and pods along with easier configuration and updates. For instructions on installing Helm, checkout the official documentation (opens new window).

If you would like more information about the DataTorch Helm Charts checkout the official charts repository (opens new window).

# Installing Chart

1. Add DataTorch repository to Helm

   helm repo add datatorch https://charts.datatorch.io
   

2. Use scripts/kub-redcred.sh to create the secret required to pull the docker images. You must have the gcrpull.json located in the same directory.
3. Create a copy of the datatorch/values.yaml and update the FQDN, database connection and licencing information.
4. Deploy an instance using helm install

   helm install my-release datatorch/datatorch -f values.yaml
   

   This will create a release with the name my-release.

# Uninstall

To uninstall/delete the my-release deployment:

helm delete my-release

The command will remove all Kubernetes components associated with the chart and deletes the release.

# SSL with NGINX and Cert-Manager

By default the DataTorch helm chart ships with the NGINX ingress. If you have disabled this chart, you many need to modify these steps to point towards your ingress.

1. Clone the official charts repository (opens new window) to access the scripts required.
2. Install cert-manager

   sh ssl/install-certmanager.sh
   

3. Create the issuers in kubernetes

   kubectl apply -f ssl/issuer-prod.yaml
   kubectl apply -f ssl/issuer-staging.yaml
   

4. Add the annotations and TLS to the ingress found in the values.yaml. Replace HOST_DOMAIN with the domain where the instance will be accessible (e.g annotator.example.com).

   ingress:
     host: HOST_DOMAIN
     annotations:
       # ... other annotations
       cert-manager.io/issuer: 'letsencrypt-prod'
       cert-manager.io/acme-challenge-type: http01
     tls:
       - secretName: datatorch-tls
         hosts:
           - HOST_DOMAIN
   

5. Install or upgrade your instance using Helm

# Azure AKS

If you have not done so already login into Azure CI:

az login

Create an Azure AKS using Azure CI:

az extension add --name aks-preview
az aks create \
   --name datatorch-cluster \
   --resource-group datatorch-resource-group \
   --node-resource-group datatorch-aks \
   --generate-ssh-keys

Checkout the AKS create command (opens new window) for more options.

Once your instance is created you can connection thought your kubectl client and deploy using the Helm instructions above.

Connect to your kubectl client using the following command:

az aks get-credentials \
  --resource-group datatorch-resource-group \
  --name datatorch-cluster

# Connection Pooling

If you are using a managed postgres instance provided by azure it is recommend to setup the microsoft pgbouncer (opens new window) to significantly reduce connection delay.

You can add the volumes and containers suggested to the backend section of the values.yaml.

backend:
  # ... other config options
  extraContainers:
    - name: azure-pgbouncer-sidecar
      image: mcr.microsoft.com/azure-oss-db-tools/pgbouncer-sidecar:latest
      # Uncomment below if you always want the latest version
      # imagePullPolicy: Always
      ports:
        - containerPort: 5432
      volumeMounts:
        - name: configfiles
          mountPath: '/etc/pgbouncer'
          # writes update the secret, we do not want to do this
          readOnly: true
      livenessProbe:
        tcpSocket:
          # This must match the port your applications use to talk to postgres
          port: 5432
        periodSeconds: 60
      lifecycle:
        preStop:
          exec:
            command: ['/bin/sh', '-c', 'killall -INT pgbouncer && sleep 120']
      securityContext:
        allowPrivilegeEscalation: false
        capabilities:
          drop: ['all']
  extraVolumes:
    - name: configfiles
      secret:
        # This must match the name of your secret above
        secretName: azure-pgbouncer-config