This release introduces powerful new capabilities for managing data replication and storage efficiency. Administrators gain enhanced visibility with new replication telemetry and write distribution metadata, while features like excluding buckets from scans and a new `CreateSession` API provide greater control over system resources and application integration.

### New Features

*   **Replication Telemetry**: Introduced telemetry for key replication operations, providing granular performance metrics and richer metadata for improved monitoring and faster troubleshooting. (#796)
*   **Exclude Buckets from Scan**: Added the ability to exclude specific buckets from background scanning, enabling administrators to reduce system load and optimize resource consumption. (#811)
*   **S3 `CreateSession` API**: Implemented the `CreateSession` API to support session-based interactions with S3 Express One Zone and standard S3 buckets, offering more flexible application integration. (#814)
*   **Write Distribution Metadata**: Objects now store 'write distribution' metadata, providing transparency into the write completeness of an object across its distributed drives to assess data resilience. (#526)
*   **Multi-Pool Support**: Enabled the configuration of multiple storage pools in multi-node deployments to facilitate more comprehensive testing of complex operational scenarios. (#587)
*   **Full Version Replication**: The `mc batch replicate` command now supports replicating all object versions, including delete markers, for more robust data protection and migration strategies. (#456)
*   **Admin API v4**: Introduced a new version of the Admin API (v4) to provide enhanced system management capabilities. (#489)
*   **Replication Diagnostics**: Added a new capability for administrators to capture detailed diagnostic information for site replication configurations to accelerate troubleshooting. (#788)

### Improvements

*   **Time-To-First-Byte (TTFB) Metrics**: Enhanced TTFB monitoring histograms to include longer time intervals, enabling more accurate performance analysis for slow operations. (#803)
*   **I/O Monitoring and Logging**: Improved the accuracy of I/O monitoring, enhanced file information retrieval, and reduced the verbosity of disk health logging for clearer diagnostics. (#781)
*   **Directory Listing and Cleanup**: Optimized memory usage when listing large directories and refined background cleanup processes for more timely storage reclamation. (#808)

### Performance

*   **Bucket Deletion**: Optimized the pre-deletion check for empty buckets, resulting in faster `DeleteBucket` operations. (#770)
*   **Data Scanner Efficiency**: Removed a significant processing delay when the data scanner encounters empty buckets, substantially reducing overall scan times in applicable environments. (#775)
*   **Asynchronous Deletes**: Optimized the deletion of upload IDs to run asynchronously, improving system responsiveness by preventing delays from network peers. (#783)

### Bug Fixes

*   **Replication Worker Configuration**: Resolved an issue where dynamic changes to the `replication_max_workers` setting were not applied without a service restart. (#785)
*   **Batch Replication Versioning**: Enforced versioning on both source and target buckets for batch replication to guarantee data consistency. **Note**: Batch replication jobs will now fail if versioning is not enabled on both buckets. (#795)
*   **Object Operations in Versioned Buckets**: Fixed an issue where `PutObject` and `DeleteObjects` could fail in versioned buckets with active lifecycle expiration rules. (#749)
*   **Health Diagnostics**: Corrected health diagnostics to prevent false positive NTP warnings, ensuring accurate time synchronization reporting. (#774)
*   **Healing Metrics**: Updated Prometheus healing metrics to exclude benign errors from dangling object cleanup, providing more accurate error counts. (#778)
*   **Read-Only License Errors**: Ensured that write operations attempted with a read-only license now report a clear error instead of failing silently. (#773)
*   **Data Write Operations**: Addressed a condition that could cause certain data write operations to freeze indefinitely and fixed a related resource leak. (#782)
*   **Policy Permissions**: Fixed a bug that incorrectly rejected the `admin:LicenseInfo` permission in policies. (#800)
*   **PostgreSQL Notifications**: Corrected the parsing of complex connection strings for PostgreSQL event notification configurations. (#798)
*   **Spark Client Recognition**: Improved recognition of Apache Spark client requests, enabling the system to apply relevant optimizations. (#801)
*   **Configuration Generation Tool**: Prevented the `minio fmt-gen` tool from failing when a console port is not explicitly provided. (#810)
*   **Internal Path Handling**: Refined the internal method for constructing paths to prevent potential issues during scanning and listing operations. (#818)
*   **User Interface Text**: Corrected a minor spelling error in user-facing text. (#710)

### Security Updates

*   **JSON Library Upgrade**: Replaced an unmaintained core JSON processing library with a modern, actively supported alternative to enhance system security and processing performance. (#786)