Skip to main content

Triage ECS Task Underutilization Check CPU_UNDERUTILIZATION

  1. Configure ECS Task Underutilization Check
  2. Use Case for ECS Task Underutilization Check
  3. Triage Guides by Violation Type
    1. Triage ECS Task Underutilization Check CPU_UNDERUTILIZATION
    2. Triage ECS Task Underutilization Check MEMORY_UNDERUTILIZATION

Understanding CPU Underutilization

Why should I care about fixing this issue?​

You should care about CPU underutilization as it suggests you may be overprovisioning resources, leading to unnecessary costs without benefiting performance.

Validate​

  • Ensure the alert is true and accurate by manually checking CloudWatch for the ECS Task's CPU metrics.
  • This policy scans reactively based on metric data obtained from CloudWatch. To manually retrieve data: look for the CPUUtilization metric for your task in CloudWatch.

Triage​

  • If left unaddressed, overprovisioning continues to incur unnecessary costs.
  • Remediating it should not cause downtime if resources are resized correctly.

Act​

  • Consider downscaling the task definition or optimizing the task's workload.

Reflect​

  • If this alert isn't useful, adjust the cpuUtilizationThreshold.