Skip to main content

Production planning

Use this checklist before production installation.

Sizing​

  • Graph Studio: size for application workload, pipelines, logs, and temporary files.
  • Graph Lakehouse: size for database nodes, query workload, frontend, config, spill, and persisted data.
  • Elasticsearch: Siemens source guidance recommends 8+ CPU cores, 64+ GiB RAM, and 100+ GiB disk for Elasticsearch servers used by Distributed Unstructured/search workloads.

Storage​

  • Configure persistent storage for product data and logs.
  • Use retention settings intentionally.
  • Plan NFS/shared storage if your deployment uses dynamic components or shared platform files.
  • Test backup and restore before production.

Networking and TLS​

  • Prefer private or restricted access paths.
  • Restrict load balancer source ranges.
  • Use organization-approved TLS certificates.
  • Use Kubernetes secrets for certificate material.
  • Restart or roll pods through the operator or Helm workflow after trust-store changes.

Licensing and secrets​

  • Prepare Altair One or release-approved license material.
  • Store credentials, license strings, tokens, keystore passwords, and private keys as secrets.
  • Do not commit real secret values to source control.
  • Confirm Graph Studio Kubernetes license/server-ID behavior for your release.

Security and OpenShift​

  • Review pod and container security contexts.
  • Confirm OpenShift security context constraints or routes when applicable.
  • Do not use deprecated privileged PodSecurityPolicy examples as modern Kubernetes guidance.

Elasticsearch​

Plan Elasticsearch if you use Distributed Unstructured or large text-search workloads. Confirm the supported Elasticsearch version range for your release before installing.