You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
The ecsobserver will discover all containers from all ECS nodes with the API. So all containers will be scraped by only 1 aws-otel-collector process.
Due to our scale, we prefer to install 1 aws-otel-collector per node in the ECS cluster that will scrape metrics only on its hosted containers.
We are compiling the aws-otel-collector with the dockerobserver for our needs.
Additional context
We currently have a cluster with ~300 nodes, ~1600 ecs services and ~7500 containers running
We plan to launch 2 aws-otel-collector per node : one for metrics and one for traces. Why 2, because in our release process and configuration management it will be easier.
This issue is created following our discussions with @alolita .
The text was updated successfully, but these errors were encountered:
Thanks for filing a request @r0mdau We will add to our backlog for addition in one of the Q2 2022 ADOT releases after the component passes evaluation, ADOT testing and security reviews.
Is your feature request related to a problem? Please describe.
The ecsobserver will discover all containers from all ECS nodes with the API. So all containers will be scraped by only 1 aws-otel-collector process.
Due to our scale, we prefer to install 1 aws-otel-collector per node in the ECS cluster that will scrape metrics only on its hosted containers.
Describe the solution you'd like
We would like to have in this aws-otel-collector the dockerobserver extension.
Describe alternatives you've considered
We are compiling the aws-otel-collector with the dockerobserver for our needs.
Additional context
We currently have a cluster with ~300 nodes, ~1600 ecs services and ~7500 containers running
We plan to launch 2 aws-otel-collector per node : one for metrics and one for traces. Why 2, because in our release process and configuration management it will be easier.
This issue is created following our discussions with @alolita .
The text was updated successfully, but these errors were encountered: