This release enhances security by changing how temporary credentials are managed and strengthens access controls for object retention. Key operational improvements include more efficient memory usage for write operations and increased flexibility for debugging distributed environments.

### Breaking Changes
*   New temporary security credentials (STS) are now derived dynamically and are no longer stored on the server to improve security. Integrations that query for newly created STS credential records will require updates, as these records are no longer persisted. (#369)

### Security Updates
*   Enforces authorization checks for setting object retention and ensures the user's access key is consistently recorded in audit logs for improved compliance. (#411)

### Improvements
*   Optimized memory management for write operations by reusing a global memory pool, reducing overall memory consumption. (#405)
*   Improves the debugging workflow for distributed environments by allowing nodes to run different binary versions and adding more detailed drive error logging. (#406)

### Bug Fixes
*   Corrects a parameter handling issue to ensure the `BatchJobCancel` operation functions as expected. (#410)
*   Ensures disabled, empty user groups are correctly synchronized during site replication healing to maintain configuration consistency. (#404)
*   Restores expected behavior for single-node, single-drive (SNSD) mode to exit immediately if the designated drive is offline at startup. (#407)
*   Resolves an issue that could cause a service interruption when scanning folders with automatic directory compaction enabled. (#413)
*   Fixes an issue where notification groups could appear empty in SNSD configurations, ensuring event notifications are properly configured. (#414)
*   Improves the accuracy of drive health assessments in distributed environments by correctly distinguishing between local and remote drives during maintenance checks. (#415)