All Products
Search
Document Center

Cloud Config:Idle GPU detection for ECS instances

Last Updated:Nov 24, 2025

An ECS instance is considered compliant if its maximum GPU utilization within a specific time range is greater than or equal to a specified value. This rule is not applicable if an ECS instance is not monitored by CloudMonitor, has no monitoring data, or is a non-GPU ECS instance. The default detection time range is the past 7 days. The detection process uses CloudMonitor data APIs and consumes the free quota for basic CloudMonitor. To ensure detection quality, you must enable Hybrid Cloud Monitoring. For more information about Hybrid Cloud Monitoring billing, see CloudMonitor billing details.

Risk level

Default risk level: Low.

You can change the risk level as needed.

Detection logic

  • An ECS instance is considered compliant if its maximum GPU utilization within a specific time range is greater than or equal to a specified value.

  • The rule is not applicable if an ECS instance is not monitored by CloudMonitor or has no monitoring data.

  • This rule does not apply to non-GPU ECS instances.

The default detection time range is the past 7 days. The detection process uses CloudMonitor data APIs and consumes the free quota for basic CloudMonitor. To ensure detection quality, you must enable Hybrid Cloud Monitoring. For more information about Hybrid Cloud Monitoring billing, see CloudMonitor billing details.

Rule details

Parameter

Description

Rule name

Idle GPU detection for ECS instances

Rule identifier

ecs-gpu-max-utilization-check

Tag

[ECS]

Automatic remediation

Not supported

Rule trigger

24-hour cycle

Supported resource types

[ACS::ECS::Instance]

Input parameters

relativeTime(Default value: 168)
utilization(Default value: 10)

Remediation guide

For remediation steps, see Elastic GPU Service (gn/vgn/sgn series).