---
title: "Performance GPU Alarms"
slug: "performance-gpu-alarms"
updated: 2026-02-23T11:38:03Z
published: 2026-02-23T11:38:03Z
---

> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.lakesidesoftware.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Performance GPU Alarms

GPU performance counters provide the relative GPU health and performance. Using the **Performance: GPU** tab, you can configure yellow and red level alarms for the GPU performance counters.

In the example alarm configuration called out below: If the GPU temperature is at or above 70° Celsius for 120 seconds a **Yellow** level alarm is triggered. If the GPU temperature is at or above 90° Celsius for 240 seconds a **Red** level alarm is triggered. If either level of alarm is triggered a **Notification** will be sent. The configured **Action** will be carried out only if a red level alarm is triggered.

> [!NOTE]
> **NOTE**: An alarm that passes through a Yellow state and achieves a Red state is considered to be in both Yellow and Red alarm status until the condition value is within the limits.

![](https://cdn.document360.io/87d11426-6b1f-4b19-b593-b71e55a81af3/Images/Documentation/alarms_performance_GPU.jpg)

The following table defines each of the GPU performance alarms.

| GPU Performance Alarm | Description |
| --- | --- |
| Fan Level | Monitors the operational health level of the GPU fan. |
| Fan RPM | Monitors the rotating speed (measured in RPM) of the GPU fan. |
| Memory Used Bytes | Monitors the number of bytes of memory the GPU card is using. |
| Memory Used Percentage | Monitors the percentage of available memory the GPU card is using. |
| Number of Applications | Monitors the number of applications currently using the GPU. |
| Power State | Monitors the power state of the graphics processing unit (GPU). A GPU can be in one of 16 power states (but not all cards support all 16 states). Values are on a scale in which 0 indicates using the most power and 15 indicates using the least power. |
| Temperature Board (°C) | Monitors the temperature (in degrees Celsius) of the GPU mother board. |
| Temperature GPU (°C) | Monitors the temperature (in degrees Celsius) of the GPU. |
| Temperature Memory (°C) | Monitors the temperature (in degrees Celsius) of the GPU memory chip. |
| Temperature Power Supply (°C) | Monitors the temperature (in degrees Celsius) of the GPU power supply. |
| Thermal Level | Monitors the health level of the GPU temperature 0 = unknown 1 = normal 2 = warning 3 = critical |
| Usage Bus | Monitors the GPU bus usage percentage (between 0 and 100%). |
| Usage Frame Buffer | Monitors the GPU frame buffer usage percentage (between 0 and 100%). The frame buffer is an area of memory used to hold the frame of data that is continuously being sent to the screen. |
| Usage GPU | Monitors the GPU usage percentage (between 0 and 100%). |
| Usage Video | Monitors the GPU video usage percentage (between 0 and 100%). |

## Configure Performance GPU Alarms

1. Use the **Performance: Alarms**checkboxes to apply the same configuration to the individual alarms.
2. To change a default **Limit** or **Time** value, click the value, then enter the updated value.
3. Depending on the alarm, selecting a check box to the left of an alarm will generate an alarm when the value is either above or below the configured **Limit** for the specified time period. For some alarms, such as **Temperature GPU**, the alarm is triggered if the value is at or *above* the Limit. For others, such as **Fan RPM,** the alarm is triggered if the value is at or *below* the Limit. If this box is unchecked, then the SysTrack Agent will ignore the limit thresholds – no alarm will be generated (although the Agent will still continue to track data for this setting).
4. Select a **Notification** profile to use for the alarm notification settings that will be followed when the alarm is triggered.
5. Select a **Time Window** for the alarm.
6. Select an **Action** profile to use when the alarm is triggered.