Skip to main content
The kubox CLI is currently in public preview, and backward compatibility isn’t guaranteed at this stage. New feature · Improvement · Bug fix · Dependency update.
v0.3.17
Maintenance release
Republished the v0.3.16 build to re-run the release pipeline. No functional changes.
v0.3.16
RBAC configuration
Define Kubernetes ClusterRoles and per-controller role bindings in cluster.yaml using the rbac and roleBinding fields. Roles are applied to the cluster automatically during creation.
v0.3.15
Helm controllers
Controllers defined as Helm charts in cluster.yaml now deploy reliably.
v0.3.14
Wildcard DNS records
Creating an AWS Route53 subdomain now also provisions a wildcard record (for example *.apps.example.com), so any subdomain routes to your cluster without extra DNS setup.
Custom resource readiness
Manifests that reference a custom resource definition that isn’t installed yet now retry automatically instead of failing. Tune this with gitOps.waitForCRD.
v0.3.13
Ingress load balancer
Expose HTTP and HTTPS traffic through the cluster load balancer by setting aws.ingress.enabled.
Route53 subdomain
Point a custom subdomain at your cluster’s load balancer automatically with dns.rootDomain and aws.route53. The DNS hosted zone is validated before the cluster is created, so misconfiguration fails early.
Controller registry
Controllers are now managed through a registry, making them easier to configure and extend.
v0.3.12
Node labels and taints
Set Kubernetes labels and taints per node group in cluster.yaml, supporting the NoSchedule, NoExecute, and PreferNoSchedule effects.
Node group configuration
nodeGroups replaces nodes in cluster.yaml. Existing configurations using nodes continue to work, and nodeGroups takes precedence when both are specified.
Node labeling reliability
Cluster creation now waits for every node to be ready, reports labeling errors instead of ignoring them, and retries on transient failures.
ECR credentials on Talos
The kubelet ECR credential provider now activates only when enable-ecr-credential-provider is set to true.
v0.3.7
S3 state backend
Restored support for storing Pulumi state in an S3 backend.
v0.3.6
This release re-architects kubox around a declarative reconcile engine, changing how kubox create works. Create or update in one command
kubox create -f cluster.yaml now reconciles your cluster to match the file. Running it again applies your changes instead of failing, so the same command handles both the first creation and later updates.
Output controls
New global flags — --debug, --trace, --quiet, --no-color, and --progress — let you control how much detail kubox prints, which is useful in CI logs.
Live progress
kubox create and kubox delete now show real-time, step-by-step progress instead of appearing to hang during long operations.
v0.3.5
Private ECR images on Talos
Talos nodes can now pull images from private Amazon ECR registries.
v0.3.4
Secret ordering
Secrets are now created earlier during cluster bootstrap, so they’re available when the components that depend on them start.
v0.3.3
AWS Secrets Manager sync
Sync secrets from AWS Secrets Manager into Kubernetes secrets, with support for generic, tls, and docker-registry types. Keys are extracted automatically from JSON secret values.
Server-side apply
Kubernetes resources are now applied using server-side apply, making repeated cluster operations idempotent and conflict-aware.
v0.3.2
Cleaner output
Removed a deprecation warning that appeared during cluster operations.
v0.3.1
Faster node boots
Cluster provisioning now uses a cached Talos image, so nodes boot faster and more reliably.
v0.3.0
Release versioning
Version bump beginning the 0.3.x release series.
v0.2.21
Talos and Go
Upgraded the default Talos version from v1.8.3 to v1.10.6, and the Go toolchain from 1.23 to 1.24.
v0.2.20
Custom resource readiness
Added a short delay between deployment stages so newly installed custom resource definitions are ready before the applications that use them are deployed.
v0.2.19
Configurable Pulumi backend
Choose where Pulumi stores cluster state: a remote S3 bucket, a local file:// path, or a custom local state directory. backendURL and localStateDir cannot be combined.
v0.2.18
Teleport authentication
Authenticate to Teleport and manage Kubernetes access directly with kubox commands that wrap the Teleport tsh client.
v0.2.17
Teleport cluster join
Clusters can securely enroll into a Teleport cluster through CI/CD, giving you auditable, token-based access to Kubox-managed clusters.
Install script
The install script is clearer and shows more information during installation.
v0.2.15
Spot instances
Fixed configuration handling for spot instances.
v0.2.14
AWS region selection
Clusters run locally now use the AWS region from your cluster configuration instead of defaulting to us-east-1.
Cluster creation visibility
Added logging while the kubox binary uploads to S3 during cluster creation.
v0.2.12
Metrics Server
Added the Kubernetes Metrics Server, enabling kubectl top and metrics-based autoscaling.
Headlamp UI
Bundled the Headlamp web UI for exploring and managing the cluster.
Environment file loading
Environment variables from .env are now loaded correctly.
v0.2.11
CLI renamed to kubox
The command-line tool is now kubox instead of kuboxctl, for a more concise command.
Action required — Update any scripts, aliases, CI jobs, or automation that call kuboxctl to use kubox.
v0.2.10
Manifest validation
kubox create now checks your Kubernetes manifests for errors before creating the cluster, so problems surface early.
AWS region source
The AWS region is now read from your cluster configuration during create and delete, regardless of your AWS CLI default region.
Credential check
kubox verifies your AWS CLI is authenticated before create and delete operations, avoiding failures from expired or missing credentials.
Cluster re-creation
You can now re-create a cluster with the same name immediately after deleting it, with no waiting period.
v0.2.8
Node labels and taints
Apply Kubernetes labels and taints to your nodes for finer scheduling control.
Trace logging
Added trace-level logging for detailed debugging. See troubleshooting.
v0.2.7
AWS IAM roles
Attach AWS IAM roles to your cluster’s EC2 instances to simplify access to AWS services. See Role for Kubox EC2 Instances.
v0.2.6
First release
First public preview release of the kubox CLI.