Hyper-V exposes hundreds of performance counters through Windows Performance Monitor (PerfMon) and Windows Management Instrumentation (WMI). These counters provide visibility into CPU, memory, storage, networking, and cluster health across your Hyper-V environment.
Knowing what each counter measures helps you identify resource contention, detect bottlenecks, and troubleshoot performance issues before they affect workloads.
This guide focuses on the most important Hyper-V performance counters. Each section explains what the counters measure, healthy ranges, and what abnormal values usually indicate. For detailed troubleshooting and best practices, follow the linked guides for each topic.
Note: The threshold values below are intended as general guidance. Every Hyper-V environment is different, so establish a performance baseline under normal operating conditions and adjust thresholds based on your workloads, hardware, and business requirements.
Hypervisor Logical Processor counters
These counters measure CPU utilization on the physical Hyper-V host. They help identify overloaded hosts and CPU contention before virtual machines begin to slow down.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V Hypervisor\Virtual Processors | Total number of virtual processors (vCPUs) assigned to all running virtual machines on the host. | Varies by environment | A high vCPU count relative to available logical processors may indicate CPU oversubscription and increased scheduling contention. |
| Hyper-V Hypervisor Logical Processor\% Total Run Time | Percentage of time each logical processor spends running workloads. | Below 70% | Host CPU contention, CPU oversubscription, or sustained workload growth. |
| Hyper-V Hypervisor Logical Processor\% Hypervisor Run Time | Percentage of CPU time consumed by the Hyper-V hypervisor itself. | Low and stable | High interrupt activity from network or storage devices, excessive VM exits caused by legacy or unoptimized drivers, or increased virtualization overhead from CPU security mitigations such as Spectre and Meltdown protections. |
These counters provide the best view of overall host CPU health. Review them alongside virtual processor counters for a complete picture.
Hypervisor Virtual Processor counters
These counters measure CPU performance from the perspective of each virtual machine.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V Hypervisor Virtual Processor\CPU Wait Time Per Dispatch | Time a virtual processor waits before receiving CPU time. | Low and stable | CPU contention, host oversubscription, oversized VMs, or NUMA issues |
| Hyper-V Hypervisor Virtual Processor\% Guest Run Time | CPU time spent running the guest operating system. | Depends on workload | CPU-intensive applications or sustained workload growth |
CPU Wait Time is often more useful than CPU utilization alone because it shows whether virtual machines are waiting for processor time even when host CPU usage appears moderate.
Dynamic Memory counters
These counters monitor how Hyper-V allocates physical memory across virtual machines.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V Dynamic Memory Balancer\Average Pressure | Overall memory pressure across the Hyper-V host. | Below 80% | Memory contention or insufficient physical RAM |
| Hyper-V Dynamic Memory Balancer\Available Memory | Physical memory available for allocation. | Above 20% available | Host memory shortage |
| Hyper-V Dynamic Memory VM\Smart Paging Working Set Size | Amount of Smart Paging used during VM startup. | Normally zero | Startup memory shortage or insufficient host RAM |
Review these counters together to understand whether Hyper-V can meet the memory demands of all running virtual machines.
Virtual Storage Device counters
These counters measure storage performance for virtual machines.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V Virtual Storage Device\Read Operations/sec | Number of storage read operations performed each second. | Depends on workload | Read-intensive workloads |
| Hyper-V Virtual Storage Device\Write Operations/sec | Number of storage write operations performed each second. | Depends on workload | Heavy write activity or backup jobs |
| PhysicalDisk\Avg. Disk sec/Transfer | Average storage latency. | Below 15 ms | Storage bottlenecks or slow disks |
Storage counters should always be interpreted together. High I/O activity is not necessarily a problem unless it is accompanied by increasing latency.
Virtual Network Adapter counters
These counters monitor traffic flowing through Hyper-V virtual network adapters.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V Virtual Network Adapter\Bytes/sec | Amount of network traffic transmitted and received. | Depends on workload | High network utilization |
| Hyper-V Virtual Network Adapter\Packets Dropped | Number of packets discarded by the virtual adapter. | Zero | Network congestion or configuration issues |
High throughput is expected for busy workloads. Packet drops are usually a stronger indicator of network problems.
Virtual Switch counters
Virtual Switch counters provide visibility into the shared virtual network used by multiple VMs. Even if an individual VM reports normal network activity, the virtual switch itself may be congested or dropping packets.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V Virtual Switch\Bytes/sec | Total network throughput across the virtual switch. | Depends on workload | Heavy traffic across multiple VMs or a busy virtual switch. |
| Hyper-V Virtual Switch\Packets Dropped/sec | Number of packets dropped by the virtual switch. | Zero | Switch-level congestion, bandwidth saturation, or virtual switch configuration issues. |
Review Virtual Switch counters alongside Virtual Network Adapter counters to determine whether the problem affects one VM or all workloads connected to the same virtual switch.
Cluster Shared Volume counters
These counters monitor storage used by Hyper-V Failover Clustering.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Cluster CSV File System\Redirected IO | Whether storage traffic is using redirected mode. | Normally zero | CSV or storage communication problems |
| Cluster CSV File System\IO Read Latency | Read latency for Cluster Shared Volumes. | <15ms | Shared storage bottlenecks |
| Cluster CSV File System\IO Write Latency | Write latency for Cluster Shared Volumes. | <15ms | Storage congestion or hardware issues |
Persistent redirected I/O often indicates an underlying storage or cluster issue that should be investigated.
Live Migration counters
These counters measure the progress and performance of Hyper-V Live Migration. They help administrators identify slow or stalled migrations before they fail.
| Counter | What it measures | Healthy range | High values usually indicate |
|---|---|---|---|
| Hyper-V VM Live Migration\Transfer Speed | Migration throughput in bytes per second. | Stable throughout the migration | A drop toward zero during an active migration may indicate network, storage, or cluster communication problems. |
| Hyper-V VM Live Migration\Migrated | Total amount of VM memory transferred during the migration. | Increases steadily until completion | If the value stops increasing, the migration may be stalled or progressing very slowly. |
These counters are most useful when viewed together. A steady increase in Migrated with consistent Transfer Speed indicates a healthy migration.
Which counters should you monitor first?
If you're building a new monitoring strategy, start with these counters.
| Area | Priority counters |
|---|---|
| CPU | % Total Run Time, CPU Wait Time Per Dispatch |
| Memory | Average Pressure, Available Memory |
| Storage | Avg. Disk sec/Transfer |
| Network | Bytes/sec, Packets Dropped |
| Cluster | Redirected IO, CSV Read/Write Latency |
Once you establish a baseline for these counters, expand monitoring to workload-specific metrics as needed.
Best practices for using Hyper-V performance counters
- Establish a baseline before creating alerts.
- Review related counters together instead of individually.
- Monitor trends rather than isolated spikes.
- Use shorter polling intervals for CPU Wait Time and Memory Pressure.
- Revisit thresholds as workloads and infrastructure change.
Monitor Hyper-V performance with ManageEngine OpManager
ManageEngine OpManager monitors your Hyper-V environment through WMI without requiring agents on Hyper-V hosts.
Key capabilities include:
- Monitor Hyper-V hosts, VMs, and clusters from one console.
- Collect CPU, memory, storage, network, and cluster counters in real time.
- Configure threshold-based alerts for key performance counters.
- View historical trends for troubleshooting and capacity planning.
- Correlate related metrics to speed up root cause analysis.
FAQs
Which Hyper-V performance counters should I monitor first?
Start with CPU Wait Time Per Dispatch, % Total Run Time, Average Pressure, Available Memory, Avg. Disk sec/Transfer, and CSV latency. These counters provide the best overall view of Hyper-V health.
What is the difference between CPU Wait Time and CPU utilization?
Are the recommended threshold values fixed?
How are Hyper-V performance counters collected?
By Ajay Sharma S
Solutions Marketer, ManageEngine