[receiver/awsecscontainermetrics] Clarify behavior of certain metric attributes

[pjanotti]

Component(s)

receiver/awsecscontainermetrics

Describe the issue you're reporting

In my experiments with the awsecscontainermetrics receiver I didn't see the attributes aws.ecs.container.finished_at and aws.ecs.container.exit_code being generated. Looking at the respective code it seems that they can be generated depending on the time of the query. I would like a confirmation if they can be generated in practice or not. The README.md should be updated to clarify the expected behavior for these attributes.

Three other attributes that called my attention are

aws.ecs.task.pull_started_at
aws.ecs.task.pull_stopped_at
aws.ecs.container.created_at

by my experiments they are fixed per ECS task and container id (for container metrics) but it will be good to have that confirmed. Example: if it is possible, depending on timing, that pull_stopped_at is not set yet while all others are then this can mean 2 distinct timeseries for many systems (at least for a while).

[github-actions]

Pinging code owners:

• receiver/awsecscontainermetrics: @Aneurysm9

See Adding Labels via Comments if you do not have permissions to add labels yourself.

[pjanotti]

Ping @Aneurysm9