Install and upgrade

  1. Before you begin
  2. Configure the values file
    1. Persistent volume size
    2. Child workload scheduling
  3. Install
    1. Verify the deployment
  4. Upgrade
  5. Next steps

Before you begin

Confirm that you have completed the following:

  1. Get started — license retrieved and image pull secret created.
  2. A platform deployment guide — Kubernetes cluster provisioned.
  3. Configure the database — PostgreSQL 15 reachable, extensions enabled.
  4. Install dependencies — cert-manager, trust-manager, Linkerd, an ingress controller, and Kueue.
  5. GPG keys — keys generated, applied, and backed up.

Configure the values file

SAS Retrieval Agent Manager is configured through a single Helm values file. The values are standardized across all supported platforms.

Tip: Use the values file generator to build a ram-values.yaml interactively. The generator runs entirely in your browser and produces the file you pass to --values below.

You can also start from the example values file and edit it by hand.

Persistent volume size

Set the storage capacity under .storage.embedding.pvc.size and .storage.application.pvc.size. The application PVC size starts at 5Gi and increases from there.

Note: The application PVC size corresponds with the amount of data purchased from SAS.

Child workload scheduling

The api.childScheduling values control the tolerations, node selectors, and affinity for workloads spawned by the API, including agents, evaluations, source pods, and model services. The example values file prefers nodes labeled for SAS Retrieval Agent Manager. Adjust these settings to match your cluster’s scheduling configuration.

Install 

helm install retrieval-agent-manager oci://ghcr.io/sassoftware/sas-retrieval-agent-manager-deployment/sas-retrieval-agent-manager \
  --version 2026.9.0 \
  --values <SAS Retrieval Agent Manager Values File> \
  -n retagentmgr \
  --create-namespace \
  --timeout 10m

Note: Use the packages section of the repository to find an installable version. If something fails and you need to redeploy, run helm uninstall retrieval-agent-manager -n retagentmgr and retry the installation.

Verify the deployment 

kubectl get pods -n retagentmgr

Upgrade

Do not redeploy or regenerate GPG keys when upgrading. Deleting the GPG secrets or configmaps during an upgrade permanently destroys all encrypted application data with no possibility of recovery. See GPG keys.

We recommend step-through upgrades, applying each version incrementally rather than skipping versions. This ensures that you get each database migration in the intended order.

You must use the same GPG keys and Helm values for upgrades as you did for the initial installation. We also recommend starting from the example values file for the version you are upgrading to and copying over any custom values from your previous file.

helm upgrade --install retrieval-agent-manager oci://ghcr.io/sassoftware/sas-retrieval-agent-manager-deployment/sas-retrieval-agent-manager \
  --version <SAS Retrieval Agent Manager Version> \
  --values <SAS Retrieval Agent Manager Values File> \
  -n retagentmgr \
  --timeout 10m

Next steps