Skip to content

System requirements

Confirm the target cluster meets the following before starting the pre-deployment checklist.

ZEN-INTEL is deployed to Kubernetes. The application services ship as container images from a container registry, and are deployed with the provided Kubernetes configuration manifests. You need:

  • A Kubernetes cluster, version 1.30 or later.
  • kubectl configured against the target cluster.
  • The CloudNativePG operator installed cluster-wide, which provides PostgreSQL. See Install the platform.
  • An ingress-nginx controller and a way to issue TLS certificates. See DNS and TLS.
  • A default storage class that supports persistent volumes, for PostgreSQL and any persistent broker or cache state.

The cluster runs the following workloads: React frontend, Go backend, three workers (ai-worker, asm-worker, brand-worker), Python AI service, PostgreSQL, RabbitMQ, and Redis.

ResourceMinimumRecommended
Control plane1 master3 masters
Bare metal nodes1 machine3 machines
vCPU (total)6 cores24 cores
Memory (total)12 GB48 GB
Storage (total)100 GB1000 GB

Size the workers and Python AI service to the expected background and AI load, since long-running jobs and AI analysis run there rather than in the API backend.

PostgreSQL, RabbitMQ, and Redis all run in-cluster and are deployed as part of the ZEN-INTEL install. There is no dependency on managed or external data services.

For data safety, PostgreSQL is backed up to external S3-compatible storage outside the cluster. Provision persistent volumes for PostgreSQL, and for RabbitMQ if durable queues are enabled, from a storage class that survives pod rescheduling.

  • Inbound: HTTPS to the ingress only.
  • Outbound: the Python AI service must reach the LLM service, and the backend, worker, and AI service must reach any enabled integrations or the Zentara Gateway. Confirm egress firewall rules before install day.