Native Out-of-the-Box (OOTB) Hyper-V CPU Contention Metrics for Hyper-V Extension

Platform
,
1

Business Problem & Description (Persona-Based)

  • Business Problem:

    1. Dynatrace currently lacks direct, native Out-of-the-Box (OOTB) metric support for Hyper-V CPU contention. Because contention must be manually derived using custom combinations of host/VM utilization and core counts, the lack of native telemetry baselines causes Davis AI to trigger a high volume of false-positive anomaly detections and alert fatigue.

    2. Operating teams are forced to maintain complex, custom DQL queries and dashboards just to approximate cluster-level contention parity available natively in other hypervisors (e.g., VMware). This creates operational overhead and degrades trust in Davis AI’s automated problem detection.

Technical Details & Current Workaround

  • Current Workaround: Manually deriving composite scores via calculations combining:

    • Real VM CPU Utilization (hyperv.virtual.machine.real.cpu.utilization)

    • Host CPU Utilization (hyperv.host.cpu.utilization)

    • vCPU Count (hyperv.virtual.machine.cpu.count)

    • Logical CPU Count (hyperv.host.logical.cpu.count)

Native Telemetry Support:

Implement native OOTB metrics for Hyper-V CPU contention and core saturation directly within the OneAgent/Extension framework, removing the need for manual metric derivation.