ZCP September Update: Kubernetes 1.37, Safer CLI Workflows, and Terraform v0.2.0

The September 6 and 7 releases bring updates across managed Kubernetes, the zcp CLI, and the Terraform / OpenTofu provider. This roundup focuses on the operational changes that affect how you create clusters, handle CLI output, and manage object-storage settings in infrastructure code.

Kubernetes 1.37 for new managed clusters

Kubernetes 1.37 is available for new ZCP managed clusters in YUL-1 and YOW-1. New clusters can use 1.37.0. The current default remains 1.36.4, and 1.35.8 and 1.34.11 remain available for new clusters.

Existing clusters do not change version automatically. Plan and validate the upgrade for your workloads, add-ons, and operating procedures. Upgrade one minor version at a time, such as 1.35 to 1.36 before moving to 1.37. The Kubernetes 1.37 announcement has the current version snapshot and portal upgrade steps.

CLI v0.0.28 and v0.0.29: safer output and clearer results

The September 7 CLI releases address security, backup workflows, connectivity, and paginated results. v0.0.28 includes the changes prepared for v0.0.27, which was never published. Upgrading to v0.0.29 includes the v0.0.28 fixes.

Most importantly, --debug output and errors built from unparseable responses now redact API tokens and credential fields. If you shared debug output from an older CLI in a ticket, chat, or repository, rotate the exposed token. This change reduces the chance of repeating that exposure. It does not invalidate a token that was already shared.

Backup operations now match the platform response and cancellation workflow. zcp backup list correctly decodes the backup-list response. zcp vm-backup delete submits an asynchronous service cancellation request, so a successful command means cancellation has begun, not that the schedule has disappeared immediately. Check the relevant list command before treating the schedule as removed.

zcp instance ssh now chooses a public IP first when one is attached, then falls back to the private IP. Use --use-public or --use-private when a script or runbook needs an explicit path. The flags are mutually exclusive.

In v0.0.29, zcp volume list retrieves every page. It now returns an error when pagination is malformed or incomplete rather than quietly returning partial or duplicate volume results. A VPC that reaches the default eight-subnet limit now receives a direct support instruction after a failed zcp network create --vpc; the CLI does not increase the quota automatically.

Use the CLI installation guide for the supported install and upgrade paths, then confirm the installed release:

zcp version

The full CLI v0.0.29 changelog records the complete release details.

Terraform provider v0.2.0

Terraform provider v0.2.0 adds resources for bucket versioning, policies, tagging, lifecycle rules, and CORS settings. It also adds VPC and multi-network support to zcp_instance, the zcp_volume data source, and root_volume and volumes on data.zcp_instance. Gateway credentials used for object-storage configuration are not exposed as resource attributes.

For example, this resource enables versioning for a bucket when the surrounding configuration already declares zcp_object_storage.assets and zcp_object_storage_bucket.media:

resource "zcp_object_storage_bucket_versioning" "media" {
  object_storage = zcp_object_storage.assets.id
  bucket         = zcp_object_storage_bucket.media.id
  enabled        = true
}

Update the provider constraint, then initialize and review the resulting plan before applying it:

terraform {
  required_providers {
    zcp = {
      source  = "zsoftly/zcp"
      version = "~> 0.2.0"
    }
  }
}

terraform init -upgrade
# Or: tofu init -upgrade

The Terraform Registry lists v0.2.0. For OpenTofu, if tofu init -upgrade reports no matching version, wait until the OpenTofu registry lists v0.2.0 and retry. Keep the provider lock-file change in review with the configuration change, especially where bucket lifecycle or policy settings are part of a production workload. Review the provider v0.2.0 changelog and the ZCP changelog for the complete resource and upgrade notes.

Update the CLI before your next maintenance task. Review the Kubernetes upgrade plan, confirm backup cancellations with a list command, and review object-storage changes in Terraform or OpenTofu before applying them.