Skip to content

Installation

Introduction

You can install Prisme.ai on any platform using Docker.
You can also follow one of our guides to install Prisme.ai on popular hosting platforms, these installations are often more production oriented.

We recommend using Kubernetes with Helm while deploying our platform.
Also, before deploying our platform, we strongly advise to familiarize yourself with it's architecture in order to understand the different microservices required as well as the different types of databases needed.
The architecture can be reviewed here.

General considerations

Provision required cloud provider resources :

  • A load balancer with ingress controller capabilities. Other load balancing methods might be used but are not documented here
  • A minimal Kubernetes cluster for Prisme.ai should have at least 3 nodes 4vCPU / 16GB, this configuration is suitable to run a smooth core installation. If your installation includes microservices from the enterprise features you may need more depending on the additional microservices.
  • Databases : Elasticsearch 7+ , Redis 5+, MongoDB 4+
    • We recommend using a managed solution, if you want to self-host those services, please check their official documentation.
  • For enterprise features :
    • a Redis with SEARCH and JSON modules is needed (included by redis/redis-stack-server docker image)
    • a RWX Kubernetes volume (50GB minimum, we do not recommend using NFS)
  • Services workspaces and runtime requires one of these document/object storage :
    • Filesystem (Kubernetes PVC that supports RWX, 50GB minimum)
    • S3 compatible object storage
    • Azure Blob storage
  • DNS :
    • Workspaces pages routing require a wildcard dns like : *.pages.prismeai.customer.com
    • API entrypoint require a dedicated DNS like : api.prismeai.customer.com
    • Studio require a dedicated DNS like : studio.prismeai.customer.com
  • Cluster machines date must remain synchronized within 10 seconds of each other

Offline environments

If the installation environment does not have access to internet, ensure that the following prerequisites are always met :

  • If using app microservices, prismeai-functions need access to a NPM registry. If the default https://registry.npmjs.org/ is not available, you must configure your own npm registry using NPM_CONFIG_REGISTRY environment variable inside prismeai-functions microservice
    • Depending on your NPM proxy, it might require additional configuration for authentication
  • If using internal / self-signed TLS certificates, they must be configured inside almost all microservices to avoid HTTPS errors
    • prismeai-searchengine and prismeai-crawler are Python microservices : once mounted, the certificate file location can be configured with REQUESTS_CA_BUNDLE
    • prismeai-llm does not need the custom certificate
    • Every other prismeai-* services are NodeJS : once mounted, the certificate file location can be configured with NODE_EXTRA_CA_CERTS

Available guides

Currently available guides :