Wondering about Prisma Horizontal Scaling setup? We can help you.

As part of our Server Management Services, we assist our customers with several Prisma queries.

Today, let us see how our Tech goes about its setup.

 

Prisma Horizontal Scaling setup

Before we begin, there are two important aspects we need to take into consideration.

• The Prisma Management API

Prisma server and Prisma CLI use this API endpoint to perform administrative tasks.

However, with this API, we can only run one instance of the Prisma service.

• The PRISMA_CONFIG

Since it contains both sensitive and non-sensitive information, it is tricky to deal with.

Moving ahead, let us see how our Support Techs perform the Horizontal Scaling setup.

• Prisma Helm Chart

The Helm chart contains two sets of Deployments and Secrets:

One that does not have the Management API and can scale up with multiple replicas and one that has the Management API and a fixed replica of 1.

• Chart folder structure

Generally, the chart contains the following files:

k8s/prisma/
├── Chart.yaml
├── README.md
├── charts
│ └── rabbitmq-2.1.1.tgz
├── requirements.lock
├── requirements.yaml
├── templates
│ ├── deployment-management.yaml
│ ├── deployment.yaml
│ ├── secret-management.yaml
│ ├── secret.yaml
│ └── service.yaml
├── secrets.enc
├── secrets.yaml
└── values.yaml

• RabbitMQ chart dependency

To get the chart dependency, we need to define requirements.yaml and follow the Helm instructions.

We can provide the configuration for the RabbitMQ chart within the main values.yaml as long as it’s grouped by a key with the same name as the dependency chart rabbitmq.

• Deployment resources

Deployments and secrets are defined twice.

However, the Docker image for Prisma production is, prismagraphql/prisma-prod. This is different from prismagraphql/prisma! (source)

• Service resource

Service resource points to the “normal” Deployment, not the management one.

• Install the chart

In order to install the chart we just have to run the helm install command:

$ helm upgrade \
--install \
--wait \
--namespace $NAMESPACE \
-f k8s/prisma/values.yaml \
-f k8s/prisma/secrets.yaml \
$HELM_RELEASE_NAME k8s/prisma

As a result, we will see something similar to this:

$ kubectl get pods
NAME READY STATUS
prisma-rabbitmq-0 1/1 Running
prisma-rabbitmq-1 1/1 Running
prisma-6cf659ccc9-8dgg4 2/2 Running
prisma-6cf659ccc9-m5fdk 2/2 Running
prisma-management-5bb959bbb5-qdbth 2/2 Running

$ kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S)
prisma-rabbitmq ClusterIP 10.35.244.85 <none> 4369/TCP,5672/TCP,25672/TCP,15672/TCP 73d
prisma ClusterIP 10.35.254.183 <none> 80/TCP

• Use the Prisma Management API

Say, we have to deploy a schema change. In such a case, we need to connect to the prisma-management-xxx-xx pod.

$ kubectl port-forward prisma-management-5bb959bbb5-qdbth 4467:4466

After that, we need to execute the Prisma deploy command:

$ prisma deploy -e .env

Here, the .env contains the following variables:

PRISMA_API_URL=http://localhost:4467/my-service/staging
PRISMA_API_SECRET=xxx
PRISMA_MANAGEMENT_API_SECRET=xxx

[Need help with the setup? We are available here for you]

 

Conclusion

In short, we saw how our Support Techs perform the Prisma Horizontal Scaling setup.