Configure Helm for common scenarios

Start your values.yaml from one of the sets below, then adjust it against the Helm Values Reference.

Prerequisites

The omnia-crds chart and the omnia release installed in omnia-system, because the apply step upgrades the release with --reuse-values. See Install with Helm.
For the production set: Istio installed in the cluster before Omnia, and a TLS Secret named agents-tls in omnia-system.
For the production set: no existing KEDA installation. If KEDA is already in the cluster, set keda.enabled: false.

Example configurations

Minimal (development)

Prometheus and Grafana only, everything else left at its default:

prometheus:
  enabled: true
grafana:
  enabled: true

Production

Adds replica count, resource limits, persistent storage for the observability stack, a TLS gateway, JWT authentication, and KEDA. The values are illustrative; size them for your load.

replicaCount: 5  # the chart default is 3; use 3 or more for HA with leader election

resources:  # the chart defaults are cpu 1, memory 512Mi limits and 100m / 256Mi requests
  limits:
    cpu: 2000m
    memory: 1Gi
  requests:
    cpu: 500m
    memory: 512Mi

prometheus:
  enabled: true
  server:
    persistentVolume:
      enabled: true
      size: 100Gi

grafana:
  enabled: true
  adminPassword: ""  # Use existingSecret instead
  admin:
    existingSecret: grafana-admin
    userKey: admin-user
    passwordKey: admin-password

loki:
  enabled: true
  singleBinary:
    persistence:
      enabled: true
      size: 50Gi

tempo:
  enabled: true
  persistence:
    enabled: true
    size: 20Gi

gateway:
  enabled: true
  listeners:
    https:
      enabled: true
      tlsSecretName: agents-tls

istio:
  enabled: true

authentication:
  enabled: true
  jwt:
    issuer: "https://auth.example.com"
    audiences:
      - "agents-api"

keda:
  enabled: true

authentication here applies Istio JWT policies to agent pods in the release namespace only. Agents in workspace namespaces need spec.externalAuth.oidc on their AgentRuntime. See Configure authentication.

See Production setup for a walkthrough that installs Istio and a Gateway alongside this values set.

Observability only

Add the observability stack to agent deployments you already manage, with the gateway, authentication, and KEDA left off:

prometheus:
  enabled: true
grafana:
  enabled: true
loki:
  enabled: true
alloy:
  enabled: true
tempo:
  enabled: true

gateway:
  enabled: false

authentication:
  enabled: false

keda:
  enabled: false

Apply the values

Save the set you chose as values.yaml.

For the production set, create the Grafana admin Secret first. Replace the placeholder password:

kubectl create secret generic grafana-admin -n omnia-system \
  --from-literal=admin-user=admin \
  --from-literal=admin-password=<password>

Upgrade the release:

helm upgrade omnia oci://ghcr.io/altairalabs/charts/omnia \
  --devel -n omnia-system --reuse-values -f values.yaml

Verify

kubectl get pods -n omnia-system

Pods matching the components you enabled (omnia-prometheus-*, omnia-grafana-*, and so on) reach Running.