Checks whether ECS GPU memory usage meets or exceeds a threshold within a specified period (default: 7 days). Instances without GPU monitoring or non-GPU instances are marked not applicable. Uses Cloud Monitor data APIs and consumes the Cloud Monitor Basic free quota. Enable Hybrid Cloud Monitoring for optimal detection (billing covered in Cloud Monitor billing details).
Risk level
Default risk level: Low.
Adjustable as needed.
Detection logic
-
Compliant: maximum GPU memory usage meets or exceeds the threshold within the specified period.
-
Not applicable: GPU not monitored by Cloud Monitor, or no monitoring data available.
-
Not applicable: non-GPU ECS instances.
Default detection period: 7 days. Uses Cloud Monitor data APIs and consumes the Cloud Monitor Basic free quota. Enable Hybrid Cloud Monitoring for optimal detection (billing covered in Cloud Monitor billing details).
Rule details
|
Parameter |
Description |
|
Rule name |
ECS GPU memory usage idle check |
|
Rule identifier |
|
|
Tag |
[ECS] |
|
Automatic remediation |
Not supported |
|
Rule trigger |
24-hour period |
|
Supported resource types |
[ACS::ECS::Instance] |
|
Input parameters |
relativeTime (Default: 168) |
Remediation guidance
Follow the remediation steps in Elastic GPU Service (gn/vgn/sgn series).