All Products
Search
Document Center

Cloud Config:ECS GPU memory usage idle check

Last Updated:Nov 21, 2025

An ECS instance is considered compliant if its maximum GPU memory usage meets or exceeds a specified threshold over a defined period. An ECS instance is considered not applicable if its GPU is not monitored by Cloud Monitor or if monitoring data is unavailable. Non-GPU ECS instances are also considered not applicable. The default detection period is the last 7 days. This check uses Cloud Monitor data APIs and consumes the free quota for Cloud Monitor Basic. For optimal detection quality, enable Hybrid Cloud Monitoring. For more information about billing for Hybrid Cloud Monitoring, see Cloud Monitor 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 memory usage meets or exceeds a specified threshold within a defined period.

  • An ECS instance is considered not applicable if its GPU is not monitored by Cloud Monitor or if monitoring data is unavailable.

  • Non-GPU ECS instances are considered not applicable.

The default detection period is the last 7 days. This check uses Cloud Monitor data APIs and consumes the free quota for Cloud Monitor Basic. For optimal detection quality, enable Hybrid Cloud Monitoring. For more information about billing for Hybrid Cloud Monitoring, see Cloud Monitor billing details.

Rule details

Parameter

Description

Rule name

ECS GPU memory usage idle check

Rule identifier

ecs-gpu-memory-max-utilization-check

Tag

[ECS]

Automatic remediation

Not supported

Rule trigger

24-hour period

Supported resource types

[ACS::ECS::Instance]

Input parameters

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

Remediation guidance

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