This release enhances operational visibility and automation with a new cluster-wide statistics API, detailed KMS metrics, and a server configuration hash to detect drift. Key updates to the batch job framework improve usability and validation, while critical fixes prevent potential data loss in tiered environments and resolve system hangs.

### New Features
- **Detect server configuration drift** with a new configuration hash in the server info endpoint, enabling automated validation and remediation. (#1324)
- **Monitor Key Management Service (KMS) health** with new Prometheus metrics that track successful and failed KMS requests for improved observability of server-side encryption. (#1335)
- **Analyze historical API usage** with a new `/minio/admin/v4/api/stats` endpoint that provides aggregated, cluster-wide API metrics for the last minute and last 24 hours. (#1323)

### Improvements
- **Generate more detailed inventory reports** using the `v2` Batch Catalog job API, which supports new object metadata fields like `IsBucketKeyEnabled` and `KmsKeyArn`. (#1317)
- **Accelerate troubleshooting of misconfigured batch jobs** with more specific and actionable error messages. (#1352)
- **Simplify recovery from disk mount ordering issues** on newer deployments using the `reorder-disks` tool, now with support for `format.bin` metadata. (#1319)
- **Ensure valid batch job configurations** with an updated `mc batch generate catalog` command that produces `v2` API templates and clarifies prefix requirements. (#1356)

### Bug Fixes
- **Restore normal operation on modern Linux kernels** by resolving a compatibility issue that caused startup or write failures on XFS filesystems with kernel 6.14+. (#1342)
- **Ensure event queue integrity** for notifications and replication by preventing the creation of empty event files during server restarts or shutdowns. (#1326)
- **Improve site-to-site replication processing** by resolving a data race condition that occurred when modifying peer cluster configurations. (#1325)
- **Receive more accurate OpenTelemetry data** with corrected trace handling and improved tracing for object listing operations. (#1316)
- **Prevent restart loops** in environments where newer MinIO versions are managed by older AIStor object-store operators. (#1332)
- **Optimize storage consumption** by automatically cleaning up temporary data left on disks after a failed object upload. (#1330)
- **Enhance self-healing operations** by correctly calculating drive offline duration, preventing unnecessary healing attempts on consistently offline drives. (#1348)
- **Restore full replication visibility** by adding the `minio_bucket_replication_received_bytes` and `_count` metrics back to the v3 metrics endpoint. (#1347)

### Breaking Changes
- **Batch job validation is now stricter**, rejecting configurations with overlapping prefixes. Job definitions must be updated to use distinct, non-overlapping prefixes to succeed. (#1333)
- **The `queuedInProgress` field in the bucket statistics API response has been renamed to `inProgress`**. Scripts parsing this JSON output must be updated to use the new field name. (#1334)