Controller Configuration

This guide covers the configuration options for the Stratos controller.

Helm Values

When deploying with Helm, configuration is passed via --set flags or a values file:

helm install stratos oci://ghcr.io/stratos-sh/charts/stratos \
  --namespace stratos-system --create-namespace \
  --set clusterName=my-cluster \
  --set syncPeriod=60s \
  --set cloudProvider=aws

Or with a values file:

values.yaml

clusterName: my-cluster
cloudProvider: aws
syncPeriod: "60s"
leaderElect: true

image:
  repository: ghcr.io/stratos-sh/stratos
  pullPolicy: IfNotPresent

replicaCount: 1

serviceAccount:
  create: true
  annotations:
    eks.amazonaws.com/role-arn: arn:aws:iam::123456789012:role/stratos-controller-role

resources:
  limits:
    cpu: 500m
    memory: 256Mi
  requests:
    cpu: 100m
    memory: 128Mi

extraEnv: []
extraArgs: []

helm install stratos oci://ghcr.io/stratos-sh/charts/stratos \
  --namespace stratos-system --create-namespace \
  -f values.yaml

Controller Flags

The Stratos controller accepts the following command-line flags. When using Helm, most of these are configured via Helm values (shown in parentheses):

Flag	Helm Value	Default	Description
--cluster-name	clusterName	""	Kubernetes cluster name. Used for cloud instance tagging. Required.
--cloud-provider	cloudProvider	aws	Cloud provider to use: aws or fake.
--sync-period	syncPeriod	30s	Minimum interval for reconciliation.
--leader-elect	leaderElect	true	Enable leader election for HA.
--metrics-bind-address	metricsBindAddress	:8080	Address for the metrics endpoint.
--health-probe-bind-address	healthProbeBindAddress	:8081	Address for health probe endpoints.

Additional flags can be passed via the extraArgs Helm value.

Zap Logger Flags

The controller uses the Zap logger with these additional flags:

Flag	Default	Description
--zap-devel	true	Development mode (human-readable output).
--zap-log-level	info	Log level: debug, info, error.
--zap-encoder	console	Log encoder: console or json.
--zap-stacktrace-level	error	Level at which to print stack traces.

To set logger flags via Helm:

helm install stratos oci://ghcr.io/stratos-sh/charts/stratos \
  --namespace stratos-system --create-namespace \
  --set clusterName=my-cluster \
  --set extraArgs[0]=--zap-encoder=json \
  --set extraArgs[1]=--zap-devel=false \
  --set extraArgs[2]=--zap-log-level=info

Production Configuration

For production environments, use JSON logging and IRSA:

values-production.yaml

clusterName: production
cloudProvider: aws
leaderElect: true

serviceAccount:
  annotations:
    eks.amazonaws.com/role-arn: arn:aws:iam::123456789012:role/stratos-controller-role

resources:
  limits:
    cpu: 500m
    memory: 256Mi
  requests:
    cpu: 100m
    memory: 128Mi

extraArgs:
  - --zap-encoder=json
  - --zap-devel=false
  - --zap-log-level=info

helm install stratos oci://ghcr.io/stratos-sh/charts/stratos \
  --namespace stratos-system --create-namespace \
  -f values-production.yaml

Health Checks

The controller exposes health endpoints:

Endpoint	Port	Description
/healthz	8081	Liveness probe - is the controller running
/readyz	8081	Readiness probe - is the controller ready to serve

These are configured automatically by the Helm chart.

Metrics

Prometheus metrics are exposed at :8080/metrics. See Monitoring for details.

Environment Variables

Variable	Description
CLUSTER_NAME	Alternative to --cluster-name flag
AWS_REGION	Default AWS region (can be overridden per NodePool)
AWS_ACCESS_KEY_ID	AWS access key (prefer IRSA instead)
AWS_SECRET_ACCESS_KEY	AWS secret key (prefer IRSA instead)

Environment variables can be passed via the extraEnv Helm value.

Next Steps

• Quickstart - Create your first NodePool
• AWS Setup - Configure AWS prerequisites