KubeAdmiral has seen continuous development with many improvements made to both its features and stability since its open-source release. It has also been battle-tested through its extensive use in our internal environments. Thus, we believe that KubeAdmiral is now production-ready and are pleased to announce the official release of version 1.0.0.

This release reflects the achievements made by the community and developers over the past year. Thanks to all the contributors who helped make this release happen!

If you're new to KubeAdmiral, feel free to browse the KubeAdmiral website, then start with KubeAdmiral Getting Start.

Feature Highlights

Unified Management of Multi-Clusters

Support for managing Kubernetes clusters from public cloud providers, private cloud providers and on-premise Kubernetes clusters.
Support for status collection of member clusters.

(Feature contributor: @gary-lgy @limhawjia @SOF3 @JackZxj )

Multi-Cluster Application Propagation

Compatibility with various types of applications, including native Kubernetes resources and Custom Resource Definitions (CRDs) .
Support for PropagationPolicy API, which provides rich semantics to help users select the appropriate target cluster, including placement, clusterSelector, clusterAffinity, tolerations, priority, and maxClusters.
Support for Duplicate and Divide scheduling mode.
Support for static weight, dynamic weight and binpack replicas strategy.
Seamless adoption of existing single-cluster resources.
Support for automatic propagation of dependencies with follower scheduling.
Support for GPU resource scheduling.

(Feature contributor: @gary-lgy @limhawjia @mrlihanbo @JackZxj @qclc @Poor12 @wy-lucky )

Scheduling Framework Enhancement

KubeAdmiral has a new extensible scheduling framework.

Support for a rich set of in-tree scheduling plugins.
Support for dynamic scheduling profiles for the scheduler.
Support for webhook scheduling plugins.
Support for configurable rescheduling triggers.

(Feature contributor: @gary-lgy @limhawjia @JackZxj @mrlihanbo @Poor12 )

Unified Federated Type

The newly introduced unified CRDs for FederatedObject and CollectedStatus, which will reduce stability risks and operation risks, and improve system performance.

(Feature contributor: @gary-lgy @limhawjia @mrlihanbo @JackZxj )

Failover

Support for auto migration, i.e. moving unschedulable pods to another cluster automatically.
Support for custom migration.
Support for NoSchedule taints.

(Feature contributor: @gary-lgy @limhawjia @Poor12 @qclc )

Status Collection for Member Cluster Resources

KubeAdmiral provides a status collection framework, and users can customize status collection fields.
KubeAdmiral provides the in-tree status aggregation plugins for Deployment, Job and Pod.

(Feature contributor: @gary-lgy @limhawjia @JackZxj )

Federated HPA

Support for federated HPA, which can be migrated from single-cluster HPA (native/custom) at low cost.

(Feature contributor: @JackZxj @wy-lucky )

Override Policy

Add OverridePolicy and ClusterOverridePolicy to define differentiated configurations when the federated resource is propagated to different clusters.
Add a set of overriders to improve the usability of override policy, including Image, Command, Args, Labels, Annotations, etc.)

(Feature contributor: @gary-lgy @qclc )