This release enhances operational insight with comprehensive distributed tracing and a new ObjectSummary API for simplified data monitoring. Key improvements to performance and data management include reduced memory consumption, more resilient server decommissioning, and a new batch job type for cataloging object metadata. These updates provide administrators with greater control, improved diagnostics, and more efficient resource utilization.

### New Features

*   **Introduced an ObjectSummary API** to retrieve high-level storage metrics, such as total object counts and sizes, simplifying monitoring and administrative tasks. (#581)
*   **Added comprehensive distributed tracing** across IAM, Replication, ILM, KMS, Batch Jobs, and other critical services for deeper operational insight and faster performance analysis. (#648)
*   **Introduced a 'catalog' batch job type** to efficiently generate filtered inventories of object metadata for data auditing and preparation. (#493)
*   **Enabled API-based management of OpenID-derived access keys**, allowing for programmatic retrieval and listing to improve automation and auditing. (#683)
*   **Implemented telemetry for RPC streams**, providing detailed tracing and metrics to improve visibility into internal system communications. (#660)
*   **Enhanced query capabilities** with separate filtering for `requestNode` and `requestHost`, enabling more precise tracking of request origins in distributed environments. (#667)
*   **Added logger webhook notifications** for license validation failures that occur during startup, ensuring administrators are promptly informed of licensing issues. (#611)

### Improvements

*   **Optimized IAM data handling** by loading it in the background, improving system responsiveness during startup and policy updates. (#626)
*   **Enhanced the server decommissioning process** to gracefully skip unreadable objects, allowing the operation to complete without interruption. (#622)
*   **Improved audit log clarity** by ensuring replicated directory deletion logs display correctly decoded names. (#690)
*   **Refined S3 Express error reporting** to return clear messages for unsupported API operations, simplifying development and debugging. (#661)
*   **Improved error log deduplication** to group unique errors by ID, resulting in cleaner logs and more efficient issue tracking. (#671)
*   **Updated internal controller and operator components** to their latest versions, incorporating functional enhancements and fixes. (#678, #682)

### Performance Improvements

*   **Re-implemented the queue storage mechanism** to significantly reduce memory consumption, especially in configurations with numerous targets. (#666)
*   **Optimized direct I/O data copy operations** with a new ringbuffer, increasing data throughput by allowing disk reads and writes to operate more independently. (#617)
*   **Improved the efficiency of internal data traversal**, reducing resource consumption during operations that scan large datasets. (#627)

### Bug Fixes

*   Resolved a startup failure that could occur when using a Postgres event target with a large number of events. (#676)
*   Ensured expired IAM tokens are properly cleared to prevent issues associated with stale credentials. (#632)
*   Prevented in-progress drive healing operations from restarting after an upgrade, ensuring faster completion. (#637)
*   Corrected the `mc replicate backlog` command to display a complete and accurate list of all replication entries. (#631)
*   Prevented potential data loss for entries at the end of high-volume buffered data streams. (#653)
*   Resolved an upgrade failure on filesystems that do not support extended attributes (xattr). (#656)
*   Fixed an issue that caused incomplete or repetitive object listings when using ILM policies. (#657)
*   Ensured queued notifications are reliably delivered once a temporarily unavailable target system is back online. (#668)
*   Restored access to Prometheus metrics for users authenticating with derived STS credentials. (#663)
*   Resolved an issue that could cause the console to display a blank page upon loading. (#664)
*   Extended the 'purge on delete' optimization to all Hadoop jobs for more efficient delete operations. (#662)
*   Ensured hotfix Docker images correctly include all required MinIO hotfix binaries. (#659)
*   Corrected error message handling to prevent unexpected behavior during remote processing operations. (#687)
*   Ceased sending webhook notifications for license failures that cause an immediate server exit, as delivery was not guaranteed. **Operational Impact**: Administrators should monitor system logs or Prometheus alerts for these specific startup failures. (#689)

### Security Updates

*   Upgraded the internal Helm component to mitigate two vulnerabilities (GO-2025-3602, GO-2025-3601) that could lead to denial-of-service. (#655)
*   Updated Go to version 1.24.2 to protect against a potential HTTP request smuggling attack (GO-2025-3563). (#638)