The biolatency, kernellatency, socketlatency probes will be removed due to their high performance overhead and relatively limited use cases, which are only necessary during instances of network jitter. In their place, a new latency diagnostic feature will be introduced, incorporating the capabilities that these probes used to offer.

To compensate for the removal of these probes, support will be extended for latency diagnostics. The new latency diagnostic functionality will provide more detailed data and a presentation that is easier to understand.

Support for custom labels will be introduced. Users will be able to add custom labels to existing metrics, enhancing the categorization and analysis of data.

Support for RDMA (Remote Direct Memory Access)-related metrics will be added, broadening the monitoring scope to include RDMA network performance.

The monitoring dashboards for pods and nodes will be separated, each being placed on its own respective page, to simplify navigation and improve the efficiency of monitoring efforts.