TiDB Operator 2.0.0 Release Notes

Release date: December 18, 2025

TiDB Operator version: 2.0.0

With the rapid development of TiDB and the Kubernetes ecosystem, TiDB Operator releases v2.0.0, which includes comprehensive refactoring from v1 to provide a more stable, efficient, and maintainable cluster management experience.

For a detailed comparison between v2 and v1, see Comparison Between TiDB Operator v2 and v1.

Major changes and improvements

Core architecture refactoring

TiDB Operator v2 includes a comprehensive redesign of the v1 architecture, with the following key changes:

• CRD splitting: split the v1 TidbCluster CRD into multiple independent CRDs for more granular component management, improving maintainability and flexibility.
• Direct Pod management: remove the dependency on StatefulSet. Pods are now managed directly, providing higher flexibility and more precise control over Pod lifecycle and scheduling behavior.
• Controller architecture upgrade: implement controller logic based on the controller-runtime framework. This simplifies controller development, improves development efficiency, and enhances system stability and reliability.

New features and enhancements

• Support the Overlay field:

  • Enable you to flexibly specify all Kubernetes-supported fields for Pods without modifying the TiDB Operator source code.
  • Provide security validation mechanisms to prevent accidental overwrites of critical system labels.

• Support topology-aware scheduling:

  • Support the EvenlySpread strategy to evenly distribute Pods across topology domains.
  • Support topology weight configuration for flexible control of instance distribution ratios across topology domains.
  • Enhance cluster high availability and fault tolerance.

• Enhance the field validation:

  • Integrate Kubernetes Validation Rules and Validating Admission Policy.
  • Support field format and value range validation.
  • Provide clear and user-friendly error messages to facilitate troubleshooting.

• Support CRD subresources:

  • Support the status subresource for unified status management.
  • Support the scale subresource to integrate with HorizontalPodAutoscaler (HPA), enabling auto-scaling.
  • Improve compatibility with the Kubernetes ecosystem.

• Support feature gates to control change behavior:

  • You can use Feature Gates to control changes that might trigger cluster node restarts, reducing the impact on cluster stability.

• Support canceling TiKV and TiFlash scale-in operations:

  • If you add new TiKV or TiFlash nodes while a scale-in operation is in progress, TiDB Operator prioritizes canceling the scale-in for nodes that are not yet fully decommissioned. This prevents the unnecessary re-creation of nodes.

Removed features

• Remove the support for Backup and Restore Based on EBS Volume Snapshots.
• Remove the tidb-scheduler component.
• Remove the following CRDs: TiDBInitializer, TiDBDashboard, DMCluster, FedVolumeBackup, FedVolumeBackupSchedule, and FedVolumeRestore.
• Remove the TiDBMonitor and TiDBNGMonitoring CRDs. Related features are integrated through other methods. For details, see Deploy Monitoring and Alerts for a TiDB Cluster.

Acknowledgments

Thanks to all the developers and community members who contributed to TiDB Operator! We look forward to your feedback and suggestions to help us improve this milestone release.