Skip to main content

Deployment & Operations

This section is for platform operators: the team that deploys and runs DynamoAI in your own Kubernetes environment with Helm. If you are looking for product usage documentation, see the DynamoGuard and DynamoEval sections; for account access and projects, see Platform & Onboarding.

The section is organized as a runbook. A new deployment reads it in order; an existing deployment jumps to the part it needs.

Deployment Path

Follow these stages in order for a new environment:

  1. Architecture: understand the install model and how the platform is put together before touching the cluster.
  2. Prepare the environment: verify everything your environment must provide, per cloud, before the deployment window starts.
  3. Deploy: install the platform with Helm.
  4. Validate: run the acceptance checks that prove the platform is operational.

Operating an Existing Deployment

  • Advanced Configuration: bring-your-own alternatives for platform dependencies, such as an external MongoDB or your own model storage.
  • Upgrades: the general upgrade procedure, plus step-by-step guides for changes tied to a version boundary.
  • Troubleshooting: known failure symptoms across installation and operations, with causes and fixes.

Conventions Used in the Guides

  • <namespace>: the Kubernetes namespace of your DynamoAI installation.
  • <your-registry>: your internal container registry that mirrors DynamoAI images.
  • Values snippets show the keys to add or change; merge them into your existing values files rather than replacing whole blocks.
  • Every requirement pairs with a verification command you can copy and run.
  • Every upgrade guide ends with a Verification and a Rollback section; read both before starting the rollout.