Hosting using Kubernetes and Helm¶

This hosting guide shows you how to self-host Prisme.ai on any platform using Kubernetes and Helm.

Under construction

This guide is currently in construction, you can expect inaccuracies and missing parts. If you find some, please do not hesitate to reach out to support@prisme.ai.

This guide assumes that you have a basic understanding of Kubernetes, Helm, and the various external services required by Prisme.ai.

Prepare the external services¶

Deploy the required external services either within your Kubernetes cluster or make sure they are accessible to the cluster if hosted externally.

MongoDB: Deploy a MongoDB instance or cluster and note down the connection string.
Elasticsearch: Set up an Elasticsearch cluster and note down the connection details.
Redis: Deploy a Redis instance and configure it for both caching and message brokering, or deploy separate instances for each purpose.
- If planning to install Enterprise features as well, an additional Redis with JSON and SEARCH modules is required, it might not be the case of your Redis intance (i.e check with redis-cli info modules). If needed, the redis/redis-stack-server image include all required modules
File Storage: Configure your chosen file storage solution among FILESYSTEM/S3_LIKE/AZURE_BLOB and obtain the necessary access credentials.

Create a cluster¶

Create a new Kubernetes cluster.
Make sure you have access to it using kubectl from your device.

Retrieve the Helm charts¶

Download the Prisme.ai Helm charts repository or add it as a Helm repo.

Option 1: Downloading the charts¶

Download the example Helm chart from the following address : https://helm.prisme.ai/charts/prismeai-core/prismeai-core-0.1.4.tgz

Option 2: Adding as a Helm repo¶

helm repo add prismeai https://helm.prisme.ai/charts
helm repo update

Initiate the chart values file :

helm show values prismeai/prismeai-core > ./values.yml

Configure values.yaml¶

On your device, modify the values.yaml to include the connection details and credentials for the external services prepared in Step 1.
It is important that each database are correctly configured, report to the environment variables table if you have any doubt.

Every required configuration field is annotated with a REQUIRED comment

Here is some general recommendations while editing values.yaml:

If default release name 'core' will be changed, update accordingly every names including "core-"
The core ingresses should be annotated with your cloud provider ingress controller annotations
All DNS, databases urls/credentials & PVC manifests should also be updated

Deploy using Helm¶

Once you are ready: using a terminal, place yourself in the folder containing the main values.yaml.
Create a namespace for Prisme.ai and install the platform using the Helm chart:

Option 1 : Charts downloaded¶

kubectl create namespace core
helm install core . --namespace core -f values.yaml

Option 2 : Added as a repo¶

kubectl create namespace core
helm install core prismeai/prismeai-core --namespace core -f values.yaml

Verify the Deployment¶

Verify that all Prisme.ai services are running correctly by checking the pod status and logs:

kubectl get pods --namespace core
kubectl logs <pod-name> --namespace core

Ensure that the services can communicate with the external services and that there are no connectivity issues.

Set Up Ingress (Optional)¶

If you want to expose Prisme.ai to the internet, set up an Ingress controller and define Ingress rules for routing traffic to the Prisme.ai services.

Ongoing Maintenance¶

Regularly back up your MongoDB and Elasticsearch data. Monitor the health of the Prisme.ai services and the external services. Apply updates to Prisme.ai and its dependencies as needed.

You have now successfully self-hosted the Prisme.ai platform on your Kubernetes cluster. Ensure that you follow best practices for security, backups, and monitoring to maintain a stable and secure environment for your applications.

Remember that this guide provides a high-level overview, and you may need to adjust the steps based on your specific environment and requirements.