Skip to content

Latest commit

 

History

History

flux-operator

Folders and files

README.md

flux-operator

Version: 0.62.1 Type: application AppVersion: v0.62.1

The Flux Operator provides a declarative API for the installation and upgrade of CNCF Flux and the ControlPlane enterprise distribution.

The operator automates the patching for hotfixes and CVEs affecting the Flux controllers container images and enables the configuration of multi-tenancy lockdown on Kubernetes and OpenShift clusters.

The operator comes with the Flux Status Page web UI exposed on port 9080 that allows users to visualize the status of the Flux installation and monitor the GitOps delivery pipeline in real-time.

Prerequisites

  • Kubernetes 1.30+
  • Helm 3.8+

Installing Flux Operator

To install the operator in the flux-system namespace:

helm install flux-operator oci://ghcr.io/controlplaneio-fluxcd/charts/flux-operator \
  --namespace flux-system \
  --create-namespace \
  --wait

Installing the Flux Web UI as a Standalone Deployment

You can install the Flux Status Web UI as a dedicated deployment in your Kubernetes cluster, separate from the Flux Operator with:

helm install flux-web oci://ghcr.io/controlplaneio-fluxcd/charts/flux-operator \
  --namespace flux-system \
  --set fullnameOverride=flux-web \
  --set web.serverOnly=true \
  --set installCRDs=false \
  --wait

To deploy the Flux controllers and to configure automated updates, see the Flux Operator documentation.

Values

Key Type Default Description
affinity object {"nodeAffinity":{"requiredDuringSchedulingIgnoredDuringExecution":{"nodeSelectorTerms":[{"matchExpressions":[{"key":"kubernetes.io/os","operator":"In","values":["linux"]}]}]}}} Pod affinity and anti-affinity settings.
apiPriority object {"enabled":false,"extraServiceAccounts":[],"level":"workload-high"} Kubernetes API priority and fairness settings.
commonAnnotations object {} Common annotations to add to all deployed objects including pods.
commonLabels object {} Common labels to add to all deployed objects including pods.
extraArgs list [] Container extra arguments.
extraEnvs list [] Container extra environment variables.
extraVolumeMounts list [] Container extra volume mounts.
extraVolumes list [] Pod extra volumes.
fullnameOverride string ""
hostNetwork bool false If true, the container ports (8080 and 8081) are exposed on the host network.
image object {"imagePullPolicy":"IfNotPresent","pullSecrets":[],"repository":"ghcr.io/controlplaneio-fluxcd/flux-operator","tag":""} Container image settings. The image tag defaults to the chart appVersion.
installCRDs bool true Install and upgrade the custom resource definitions.
livenessProbe object {"httpGet":{"path":"/healthz","port":8081},"initialDelaySeconds":15,"periodSeconds":20} Container liveness probe settings.
logLevel string "info" Container logging level flag.
marketplace object {"account":"","license":"","type":""} Marketplace settings.
multitenancy object {"defaultServiceAccount":"flux-operator","defaultWorkloadIdentityServiceAccount":"flux-operator","enabled":false,"enabledForWorkloadIdentity":false} Enable multitenancy lockdown for the ResourceSet APIs.
nameOverride string ""
nodeSelector object {} Pod Node Selector settings.
podSecurityContext object {} Pod security context settings.
priorityClassName string "" Pod priority class name. Recommended value is system-cluster-critical.
rbac.create bool true Grant the cluster-admin role to the flux-operator service account (required for the Flux Instance deployment).
rbac.createAggregation bool true Grant the Kubernetes view, edit and admin roles access to ResourceSet APIs.
readinessProbe object {"httpGet":{"path":"/readyz","port":8081},"initialDelaySeconds":5,"periodSeconds":10} Container readiness probe settings.
reporting object {"interval":"5m"} Flux reporting settings.
resources object {"limits":{"cpu":"2000m","memory":"1Gi"},"requests":{"cpu":"100m","memory":"64Mi"}} Container resources requests and limits settings.
securityContext object {"allowPrivilegeEscalation":false,"capabilities":{"drop":["ALL"]},"readOnlyRootFilesystem":true,"runAsNonRoot":true,"seccompProfile":{"type":"RuntimeDefault"}} Container security context settings. The default is compliant with the pod security restricted profile.
service.ipFamilyPolicy string "" Sets the IP family policy on all Service resources. Uses Kubernetes defaults if unset
serviceAccount object {"automount":true,"create":true,"name":""} Pod service account settings. The name of the service account defaults to the release name.
serviceMonitor object {"create":false,"interval":"60s","labels":{},"scrapeTimeout":"30s"} Prometheus Operator scraping settings.
spiffe object {"brokerAPI":{"endpoint":"","ids":[]},"workloadAPI":{"csiDriver":"","endpoint":""}} SPIFFE settings for the flux-operator pod.
spiffe.brokerAPI.endpoint string "" The SPIFFE Broker API endpoint. Sets the --spiffe-broker-endpoint flag. Requires spiffe.workloadAPI.endpoint and spiffe.brokerAPI.ids.
spiffe.brokerAPI.ids list [] SPIFFE IDs or prefixes of the Broker API endpoint. Sets the repeatable --spiffe-broker-id flag.
spiffe.workloadAPI.csiDriver string "" The SPIFFE CSI driver name used to mount the Workload API socket at /spiffe-workload-api. When set, the endpoint must be a socket under that path, e.g. unix:///spiffe-workload-api/spire-agent.sock.
spiffe.workloadAPI.endpoint string "" The SPIFFE Workload API endpoint. Sets the SPIFFE_ENDPOINT_SOCKET environment variable.
strategy object {} Deployment update strategy. Set {"type":"Recreate"} for single-replica installs whose web UI is fronted by a load balancer health check, where a rolling update would otherwise deadlock.
tolerations list [] Pod tolerations settings.
web.authorizedClientSPIFFEIDs list [] SPIFFE IDs or prefixes of authorized web UI clients. When set, the web server on port 9080 accepts only SPIFFE mTLS connections from these clients, so the Ingress or Gateway in front of it must present an SVID (e.g. Envoy Gateway with a BackendTLSPolicy). Requires spiffe.workloadAPI.endpoint.
web.config object {} The spec of the Web Config API
web.configSecretName string "" Reference to an existing Secret in the same namespace as the deployment containing the Web Config API. Should have the key config.yaml.
web.enabled bool true Enable the Flux Status web server on port 9080.
web.httpRoute object {"annotations":{},"enabled":false,"hostnames":[],"parentRefs":[]} Gateway API HTTPRoute settings for the Flux Status web interface.
web.ingress object {"annotations":{},"className":"","enabled":false,"hosts":[],"tls":[]} Ingress settings for the Flux Status web interface.
web.networkPolicy object {"create":true} Create a NetworkPolicy to allow access to the Flux Status web interface.
web.rbac object {"createAggregation":false,"createRoles":true} Create standard roles for user access management.
web.serverOnly bool false Run the Flux Status web server as a standalone deployment (requires a dedicated Helm release).
web.serverReplicas int 1 Number of replicas for the Flux Status web server standalone deployment (only applicable if web.serverOnly is true).
web.userActions object {"access":"Impersonated"} GitOps actions configuration for the web UI.
web.userActions.access string "Impersonated" Access mode for GitOps actions. When set to FineGrained, actions are performed using the web server's own privileges instead of impersonating the user, and the web ClusterRole is extended with the native Kubernetes permissions required by the actions (only effective in web.serverOnly mode; otherwise the operator already runs as cluster-admin). Either Impersonated (default) or FineGrained.

Source Code