This release enhances performance diagnostics and operational efficiency. Administrators can now more rapidly identify slow requests using Time To First Byte (TTFB) trace filtering and perform zero-downtime license updates to maintain continuous service availability.

### Improvements

*   **Enabled zero-downtime license updates** by allowing administrators to reload a new license via a `SIGHUP` signal, eliminating the need for a server restart. The server provides immediate log feedback on the new license's validity. (#1221)
*   **Added Time To First Byte (TTFB) filtering** for server-side traces, enabling administrators to more rapidly diagnose performance bottlenecks by isolating slow-responding requests. (#1227)

### Bug Fixes

*   **Resolved a regression that prevented successful authentication** using Kubernetes or Certificate-based Single Sign-On (SSO), restoring expected login behavior for these methods. (#1230)
*   **Corrected an issue where Information Lifecycle Management (ILM) scan metrics** were incorrectly reported as zero in Prometheus, providing accurate visibility into ILM policy activity. (#1231)
*   **Fixed an error that prevented the successful disabling** of the last replication rule on a bucket, ensuring administrators can fully manage replication configurations. (#1202)
*   **Optimized resource consumption by reducing the default number** of background healing worker threads. This change prevents excessive CPU and memory usage on systems with a high core count. (#1223)