Table of Contents
Global
| Key | Type | Default | Description |
|---|
global.imageRegistry | string | "" | Override the registry host for ALL images. When set, replaces the per-image .registry. Use for air-gapped mirror deployments. |
global.imagePullSecrets | list | [] | Image pull Secrets used by every pod (names of pre-created dockerconfigjson Secrets). Defaults point at AWS ECR (a private registry), so you need one here — either the Secret ecrPullSecret manages for you, or your own if you're mirroring into a private registry instead. |
nameOverride | string | "" | Override the chart name (part of resource names). |
fullnameOverride | string | "" | Fully override the generated resource name prefix. |
commonLabels | object | {} | Extra labels added to every resource. |
commonAnnotations | object | {} | Extra annotations added to every resource. |
Pod Security Context
Applied at the pod level across all components. Compatible with restricted Pod Security Admission and OpenShift SCCs.
| Key | Type | Default | Description |
|---|
podSecurityContext.enabled | bool | true | Set to false to omit the pod security context entirely. |
podSecurityContext.runAsNonRoot | bool | true | Require the container to run as a non-root user. |
podSecurityContext.runAsUser | int | 65532 | UID to run containers as. |
podSecurityContext.runAsGroup | int | 65532 | GID to run containers as. |
podSecurityContext.fsGroup | int | 65532 | fsGroup for volume ownership. |
podSecurityContext.seccompProfile.type | string | RuntimeDefault | Seccomp profile type. |
Container Security Context
Applied to all app containers (licensed, dashboard, activityd).
| Key | Type | Default | Description |
|---|
containerSecurityContext.enabled | bool | true | Toggle the container security context. |
containerSecurityContext.runAsNonRoot | bool | true | Run the container as a non-root user. |
containerSecurityContext.runAsUser | int | 65532 | Container UID. |
containerSecurityContext.runAsGroup | int | 65532 | Container GID. |
containerSecurityContext.allowPrivilegeEscalation | bool | false | Prevent privilege escalation. |
containerSecurityContext.readOnlyRootFilesystem | bool | true | Mount the root filesystem read-only. |
containerSecurityContext.privileged | bool | false | Disable privileged mode. |
containerSecurityContext.capabilities.drop | list | ["ALL"] | Drop all Linux capabilities. |
containerSecurityContext.seccompProfile.type | string | RuntimeDefault | Seccomp profile type. |
Service Account & RBAC
| Key | Type | Default | Description |
|---|
serviceAccount.create | bool | true | Create ServiceAccounts for chart components. |
serviceAccount.annotations | object | {} | Annotations for the created ServiceAccounts (e.g. for IRSA). |
rbac.create | bool | true | Create least-privilege Roles and RoleBindings. The keygen Job's ServiceAccount may only create the delegate Secret/ConfigMap; the dashboard may only patch the licensed Deployment. |
licensed
The stateless seat arbitrator — the core service that tenon CLI clients talk to for checkout,
heartbeat, and checkin.
| Key | Type | Default | Description |
|---|
licensed.image.registry | string | 122628694949.dkr.ecr-fips.us-gov-west-1.amazonaws.com | Image registry. Defaults to the AWS ECR registry Tangram Flex publishes to — not the internal GitLab registry. Override this (or global.imageRegistry, to retarget every image at once) to pull from your own private registry/mirror. |
licensed.image.repository | string | tangram/tenon/license-manager/licensed | Image repository path. |
licensed.image.tag | string | "0.3.0" | Image tag. Must be pinned explicitly to a release tag. |
licensed.image.pullPolicy | string | IfNotPresent | Image pull policy. |
licensed.replicaCount | int | 2 | Number of licensed replicas. Stateless — safe to scale for HA. |
licensed.leaseTTL | string | 4h | Floating (shared-pool) lease time-to-live. A dead client's seat frees itself after this. |
licensed.nodeLockTTL | string | 2160h | Node-locked lease TTL (machine-bound seats). |
licensed.service.type | string | ClusterIP | Service type. Keep ClusterIP for production; clients on the enclave LAN reach it directly, or expose it another way appropriate to your network. |
licensed.service.port | int | 8080 | Service port. |
licensed.service.nodePort | string | "" | NodePort (only when type=NodePort). Empty = auto-assign. |
licensed.resources.requests.cpu | string | 25m | CPU request. |
licensed.resources.requests.memory | string | 32Mi | Memory request. |
licensed.resources.limits.memory | string | 128Mi | Memory limit. |
licensed.podDisruptionBudget.enabled | bool | true | Create a PodDisruptionBudget for licensed. |
licensed.podDisruptionBudget.minAvailable | int | 1 | Minimum available licensed pods during voluntary disruptions. |
licensed.autoscaling.enabled | bool | false | Enable a HorizontalPodAutoscaler for licensed. |
licensed.autoscaling.minReplicas | int | 2 | HPA minimum replicas. |
licensed.autoscaling.maxReplicas | int | 6 | HPA maximum replicas. |
licensed.autoscaling.targetCPUUtilizationPercentage | int | 80 | HPA CPU utilization target. |
licensed.fips | bool | false | Enable Go FIPS 140-3 mode at runtime (requires a FIPS-built image). |
licensed.extraEnv | list | [] | Extra environment variables for licensed (list of {name, value}). |
licensed.nodeSelector | object | {} | Node selector for licensed pods. |
licensed.tolerations | list | [] | Tolerations for licensed pods. |
licensed.affinity | object | {} | Affinity rules for licensed pods. |
adminToken
The shared admin credential for licensed's protected read APIs and for activityd.
| Key | Type | Default | Description |
|---|
adminToken.existingSecret | string | "" | Use an existing Secret instead of generating one. The Secret must contain key token. |
activity
Controls the Action Events subsystem: activityd and its SQLite activity store.
| Key | Type | Default | Description |
|---|
activity.enabled | bool | true | Enable Action Events. When false, no activityd is deployed, POST /v1/actions accepts and discards, and activity reads return 404. |
activity.retentionDays | int | 365 | Events older than this are deleted. Also the window in which an expired lease may still read its own history. |
activity.lateAcceptDays | int | 7 | How long after a lease expires its offline-spooled actions are still accepted. |
activity.maxActionsPerLeasePerHour | int | 1000 | Abuse guard: maximum accepted actions per lease per clock hour. |
activity.bufferMaxLen | int | 500000 | Hard cap on the Valkey ingest buffer if activityd is down. |
activity.logToStdout | bool | true | Log every event as a structured JSON line on licensed's stdout (for SIEM integration). |
activity.persistence.size | string | 10Gi | activityd database PVC size. (~5 MB/day for 50 users × 200 runs.) |
activity.persistence.storageClass | string | "" | StorageClass for the data PVC. Empty = cluster default. |
activity.persistence.accessMode | string | ReadWriteOnce | PVC access mode. |
activity.backup.enabled | bool | true | Take online SQLite snapshots. |
activity.backup.interval | string | 24h | Snapshot interval (Go duration). |
activity.backup.keep | int | 14 | Number of newest snapshots to keep. |
activity.backup.existingClaim | string | "" | Use an existing PVC for snapshots. Empty = chart creates one. |
activity.backup.size | string | 10Gi | Size of the chart-created backup PVC. |
activity.backup.storageClass | string | "" | StorageClass for the backup PVC. Empty = cluster default. |
activity.activityd.resources.requests.cpu | string | 25m | CPU request for activityd. |
activity.activityd.resources.requests.memory | string | 64Mi | Memory request for activityd. |
activity.activityd.resources.limits.memory | string | 256Mi | Memory limit for activityd. |
activity.activityd.nodeSelector | object | {} | Node selector for the activityd pod. |
activity.activityd.tolerations | list | [] | Tolerations for the activityd pod. |
activity.activityd.affinity | object | {} | Affinity for the activityd pod. |
activity.activityd.extraEnv | list | [] | Extra environment variables for activityd. |
dashboard
The operator web UI for viewing seat status, pool controls, and activity events.
| Key | Type | Default | Description |
|---|
dashboard.enabled | bool | true | Deploy the operator dashboard. |
dashboard.image.registry | string | 122628694949.dkr.ecr-fips.us-gov-west-1.amazonaws.com | Image registry. Same AWS ECR default as licensed.image.registry above. |
dashboard.image.repository | string | tangram/tenon/license-manager/dashboard | Image repository path. |
dashboard.image.tag | string | "0.3.0" | Image tag. Must be pinned explicitly. |
dashboard.image.pullPolicy | string | IfNotPresent | Image pull policy. |
dashboard.replicaCount | int | 1 | Dashboard replica count. |
dashboard.service.type | string | ClusterIP | Service type. Expose via ingress or port-forward. |
dashboard.service.port | int | 8080 | Service port. |
dashboard.ingress.enabled | bool | false | Create a Kubernetes Ingress for the dashboard. |
dashboard.ingress.className | string | "" | IngressClass to use (e.g. nginx). Empty = cluster default IngressClass. |
dashboard.ingress.annotations | object | {} | Ingress annotations (e.g. cert-manager, nginx). |
dashboard.ingress.hosts | list | (see values.yaml) | Hostnames and paths routed to the dashboard. |
dashboard.ingress.tls | list | [] | TLS blocks. Reference an existing Secret or one cert-manager creates. |
dashboard.resources.requests.cpu | string | 25m | CPU request. |
dashboard.resources.requests.memory | string | 32Mi | Memory request. |
dashboard.resources.limits.memory | string | 128Mi | Memory limit. |
dashboard.nodeSelector | object | {} | Node selector for dashboard pods. |
dashboard.tolerations | list | [] | Tolerations for dashboard pods. |
dashboard.affinity | object | {} | Affinity for dashboard pods. |
dashboard.auth.enabled | bool | true | Require admin sign-in for mutating actions (Release-all, Restart). Viewing is always read-only without sign-in. |
dashboard.auth.existingSecret | string | "" | Use an existing Secret instead of generating one (keys: admin-user, admin-password). |
dashboard.auth.adminUser | string | admin | Admin username. |
valkey
Bundled single-node Valkey (Redis-compatible) for seat-state storage.
| Key | Type | Default | Description |
|---|
valkey.enabled | bool | true | Deploy a bundled single-node Valkey. Set false to use an external instance. |
valkey.externalAddr | string | "" | External Valkey address (host:port) when enabled: false. |
valkey.existingPasswordSecret | string | "" | Name of a pre-created Secret holding a Valkey password (enables AUTH). |
valkey.existingPasswordSecretKey | string | password | Key within the password Secret. |
valkey.image.registry | string | docker.io | Valkey image registry. |
valkey.image.repository | string | valkey/valkey | Valkey image repository. |
valkey.image.tag | string | 8-alpine | Valkey image tag. |
valkey.image.pullPolicy | string | IfNotPresent | Image pull policy. |
valkey.fsGroup | int | 999 | fsGroup for the Valkey pod (matches the valkey uid/gid in the official image). |
valkey.persistence.enabled | bool | true | Persist seat state (AOF) on a PVC so seats survive a Valkey restart. |
valkey.persistence.size | string | 1Gi | PVC size. |
valkey.persistence.storageClass | string | "" | StorageClass. Empty = cluster default CSI. |
valkey.persistence.accessMode | string | ReadWriteOnce | PVC access mode. |
valkey.resources.requests.cpu | string | 50m | CPU request. |
valkey.resources.requests.memory | string | 64Mi | Memory request. |
valkey.resources.limits.memory | string | 256Mi | Memory limit. |
keygen
The Helm pre-install/pre-upgrade hook that generates the in-enclave delegate keypair.
| Key | Type | Default | Description |
|---|
keygen.enabled | bool | true | Run the in-enclave keygen Job. Idempotent — never rotates an existing key. |
keygen.delegateSecret | string | licensed-delegate | Secret the keygen Job writes the delegate private key into. Never leaves the enclave. |
keygen.delegatePubConfigMap | string | licensed-delegate-pub | ConfigMap the keygen Job writes the delegate public key into (for the operator to KMS-sign). |
keygen.backoffLimit | int | 3 | Kubernetes Job backoffLimit. |
trust
The KMS-signed delegation and KMS root public key applied by the operator after keygen.
| Key | Type | Default | Description |
|---|
trust.secretName | string | licensed-trust | Name of the Secret holding delegation.json + root-pub.pem. Normally applied out-of-band after KMS signing. |
trust.waitForTrust | bool | true | Block licensed from starting until the trust Secret exists. When false, licensed will crash-loop until trust is present. |
trust.delegation | string | "" | Inline delegation.json content. If set (e.g. --set-file trust.delegation=delegation.json), the chart creates the trust Secret for you (single-step install). |
trust.rootPublicKey | string | "" | Inline root-pub.pem content. Pair with trust.delegation. |
ecrPullSecret
A CronJob that periodically refreshes a Kubernetes docker-registry Secret from AWS ECR. ECR tokens
expire every 12 hours; this job fetches a fresh token on a schedule and writes it into a named Secret
that you reference in global.imagePullSecrets. Enable this when your images are hosted in AWS ECR
(e.g. in a GovCloud air-gapped enclave).
| Key | Type | Default | Description |
|---|
ecrPullSecret.enabled | bool | false | Enable the ECR token-refresh CronJob and its supporting RBAC, Secret, and ConfigMap. |
ecrPullSecret.secretName | string | tangram-ecr-registry | Name of the docker-registry Secret the CronJob creates/refreshes. Add this name to global.imagePullSecrets. |
ecrPullSecret.registryHost | string | "" | Full ECR registry host, e.g. <account>.dkr.ecr-fips.<region>.amazonaws.com. Empty (the default) derives it from awsAccount/awsRegion using the FIPS endpoint pattern. Override for a non-FIPS endpoint or any other custom ECR hostname. |
ecrPullSecret.awsAccount | string | "" | AWS account ID (numeric, e.g. "123456789012"). Used to derive the default registryHost (FIPS pattern) when registryHost is empty. Required when enabled: true and registryHost is not set. |
ecrPullSecret.awsRegion | string | us-east-1 | AWS region where the ECR registry lives. |
ecrPullSecret.schedule | string | "0 */10 * * *" | CronJob schedule. ECR tokens expire every 12 h; refreshing every 10 h adds margin. |
ecrPullSecret.existingSecret | string | "" | Name of a pre-existing Secret containing AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY. When set, the chart does not create a credentials Secret and accessKeyId/secretAccessKey are ignored. |
ecrPullSecret.accessKeyId | string | "" | AWS Access Key ID. Ignored when existingSecret is set. |
ecrPullSecret.secretAccessKey | string | "" | AWS Secret Access Key. Ignored when existingSecret is set. |
ecrPullSecret.awsCliImage | object | {} | Image overrides for the init container that calls aws ecr get-login-password. Sub-fields: registry, repository, tag, pullPolicy. Defaults to docker.io/amazon/aws-cli:latest with pullPolicy: IfNotPresent (this CronJob runs every few hours, often in restricted-network enclaves). |
ecrPullSecret.kubectlImage | object | {} | Image overrides for the main container that runs kubectl to create the docker-registry Secret. Sub-fields: registry, repository, tag, pullPolicy. Defaults to docker.io/bitnami/kubectl:latest with pullPolicy: IfNotPresent. |
ecrPullSecret.imagePullSecrets | list | [] | Pull secrets for the CronJob pod's own images (i.e. awsCliImage and kubectlImage). Use when those images are mirrored in a private registry separate from ECR. |
networkPolicy
Default-deny + explicit-allow NetworkPolicies (recommended for enclaves).
| Key | Type | Default | Description |
|---|
networkPolicy.enabled | bool | true | Install default-deny + explicit-allow NetworkPolicies. |
networkPolicy.clientNamespaces | list | [] | Namespaces (by name) whose pods may reach licensed:8080. Empty = same namespace only. |
networkPolicy.clientCIDRs | list | [] | Additional ingress CIDRs allowed to reach licensed:8080 (e.g. the enclave LAN). |
networkPolicy.dashboardFromNamespaces | list | [] | Namespaces allowed to reach the dashboard on port 8080. Set to your ingress controller namespace (e.g. ingress-nginx) when exposing the dashboard via ingress under a strict CNI. |