Changes and Releases
Updates on Giant Swarm workload cluster releases, apps, UI improvements and documentation changes.
Changed
- Use CoreDNS as upstream in order to avoid double configuration for internal domains.
- Internal change: add changes & releases categories for Cluster API cluster apps (#1601)
Added
- Add a check if VPA capabilities exist in the first place.
Changed
- Upgrade ATS to
v0.2.9 - Bump upstream flux toolkit version to from
v0.35.0 to v0.36.0. There are
no breaking changes. - Change resource requests & limits.
- Add weekly highlights for 2022-10-27 (#1595)
- Remove Rest API and gsctl from changes and releases (#1600)
Highlights for the week ending October 27, 2022
Feedback wanted
- We are currently designing the kubectl-gs CLI command for deleting workload clusters. The specification is available in a public issue. We’d love to see you add comments, ideas, and expectations.
- Our survey regarding cluster creation is still open. If your organization’s perspective isn’t included yet, you still have some time.
Apps
- kong-app version v2.13.0 Update kong/kubernetes-ingress-controller to 2.7.0 and require at least Kubernetes 1.22.
- linkerd2-cni version v0.8.0 Update linkerd2-cni to stable-2.12.2. In this release the app has been renamed as
linkerd2-cni, dropping the -app suffix. - linkerd-control-plane version v0.8.0 Update linkerd to stable-2.12.2. In this release the app has been renamed as
linkerd-control-plane following upstream’s naming convention. - kyverno-app version v0.11.6 updates policy-reporter version 2.10.1.
- security-pack version 0.8.1 includes the new versions of Kyverno and starboard-exporter described in this announcement.
- starboard-exporter version v0.6.2 makes the exporter’s
ServiceMonitor relabelings user-configurable, enabling custom metric collection and aggregation use cases.
Changed
- Rename app as
linkerd-control-plane (#117). - Bump version 2.12.2 (#116).
- Extend readonly ClusterRole to include pods/log so customers can debug failures on the management clusters.
- Extend readonly ClusterRole to include pods/log so customers can debug failures on the management clusters.