[metrics]: ECS average CpuUtilized reporting sum numbers instead of average

[deggja]

Context information

• AWS service: ECS
• CloudWatch namespace: ECS/ContainerInsights
• Link to metrics documentation for this service: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/cloudwatch-metrics.html
• AWS region of the exporter: Exporter not running in AWS
• AWS region of the service: eu-north-1

Exporter configuration

 region: eu-north-1
    role_arn: redacted
    metrics:
    # ECS
    - aws_namespace: ECS/ContainerInsights
      aws_metric_name:  CpuUtilized
      aws_dimensions: [ClusterName, ServiceName]
      aws_statistics: [Average]
    - aws_namespace: ECS/ContainerInsights
      aws_metric_name:  MemoryUtilized
      aws_dimensions: [ClusterName, ServiceName]
      aws_statistics: [Average] 

Exporter logs

No logs when starting the exporter with this flag

What do you expect to happen?

I expect to see values between 0-100 as this metric is supposed to get the average value for cpu usage in percent.

What happened instead?

The exporter seems to be pulling the aws_statistic for [Sum] instead of [Average]. Even though the exporter is reporting using the Average statistic, the numbers I'm seeing match the Sum statistic in AWS console.

I have not used this exporter a whole lot, but it seems to work fine on other metrics. Any pointers in the right direction highly appreciated!

Thanks.

[matthiasr]

Interesting and odd! 😄 No idea what's going on here, if anyone knows, please do tell!

[deggja]

Hey @matthiasr,

I've updated my manifests using the latest version of the exporter and the numbers are adding up. As far as I can see, this bug is no longer present.

Thanks 🚀