# User input delay

Determining performance degradations based solely on standard system health metrics like CPU, memory, and disk input or output is a difficult task. For example, on virtual desktops, the standard metrics do not often correlate with perceived performance.

The Nexthink solution measures the responsiveness of a Windows session by monitoring the user input delay per session, a Windows performance counter. That counter measures how long any user input using a mouse or a keyboard waits before a process picks it. That value is available for local and remote sessions.

### Configuration <a href="#userinputdelay-configuration" id="userinputdelay-configuration"></a>

By default, Collector reports the user input delay values as part of the user activity metrics. You can stop monitoring that by disabling the user activity reporting. Find the procedure on the [Configuring Collector level anonymization](https://docs.nexthink.com/platform/configuring_nexthink/bringing-data-into-your-nexthink-instance/deploying-nexthink-in-non-vdi-environment/installing-collector/configuring-collector-level-anonymization) page.

### Supported operating systems <a href="#userinputdelay-supportedoperatingsystems" id="userinputdelay-supportedoperatingsystems"></a>

All supported Windows versions:

* Windows 11
* Windows Server 2022

### NQL example <a href="#userinputdelay-nqlexample" id="userinputdelay-nqlexample"></a>

```
devices during past 24h
| with session.events
| compute
    medium_ratio = duration_with_medium_user_input_delay.sum()/user_interaction_time.sum(),
    high_ratio = duration_with_high_user_input_delay.sum() / user_interaction_time.sum()
| where medium_ratio > 0.10 or high_ratio > 0.05
| list device.name, medium_ratio, high_ratio
```