How to get only the average system metrics from Cloudwatch to Elasticsearch?

Current setup: Cloudwatch collecting metrics from different namespaces and the data gets ingested to Elasticsearch. With the below configuration (i.e AWS/EC2 namespace, metric name: CPUUtilization) Cloudwatch is getting and ingesting all the metrics: min, max, average..etc.


- module: aws
  period: 300s
    - cloudwatch
    - namespace: AWS/EBS
    - namespace: AWS/EC2

I followed the instruction here to only ingest the average CPUUtilization, after deploying the code below, elasticsearch still shows all the metrics.

Here's what i have tried:

- namespace: AWS/EC2
  name: ["CPUUtilization"]
  statistics: ["Average"]

and also this:

- namespace: AWS/EC2
  resource_type: ec2:instance
  name: ["CPUUtilization"]
  statistics: ["Average"]
    - name: InstanceId
      value: instanceId1, instanceId2, instanceId3#also tried only one instance ID here.  

Also it is worth noting that we run the metricbeat node on the ECS Fargate.