Feature Request

Opening a feature request kicks off a discussion.

Proposal: Update the AWS CloudWatch Input plugin to leverage the Active Metric Filter feature.

Current behavior: The ListMetric API queries for all metrics that exist in the last 2 weeks and return that data. For AWS usage that is extremely active, this can represent a lot of data to query. Most use-cases of Telegraf are only looking for data from the last few minutes to the last day in their GetMetricData calls.

Desired behavior: Apply a time filter, which would align with the GetMetricData's metricQuery windowEnd/windowStart time assignments, defined by the "delay" parameter in the config file. This will only query the ListMetric API for data that would be valuable.

Use case:

• For global scale requests, round-trip time spent per sequential API call can jeopardize the capability of returning all data from the CloudWatch APIs within a given interval of Telegraf.
• For AWS use cases that have a dense amount of data to query from, minimizing API calls with throw-away data presents cost savings (https://aws.amazon.com/cloudwatch/pricing/)
• For AWS use cases which have to share the ListMetric API with other services, this will reduce the usage and minimize footprint on the API throttling limits. ( https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/cloudwatch_limits.html)

This new feature is scheduled for Go-Live July 24th 2020. Documents to the new API changes to be posted at go-live.