ECSDog is a standalone Go application that scrapes metrics and events from ECS, and sends them to statsd.
The recommended usage is to run it with Docker (e.g. in an ECS service):
$ docker run -e AWS_REGION=us-east-1 remind101/ecsdog -cluster $CLUSTER -statsd "statsd:8125"
aws.ecs.services{cluster}
aws.ecs.service.deployments{cluster,service}
aws.ecs.service.desired{cluster,service}
aws.ecs.service.pending{cluster,service}
aws.ecs.service.running{cluster,service}
aws.ecs.service.deployments{cluster,service}
aws.ecs.service.deployment.desired{cluster,service,deployment}
aws.ecs.service.deployment.pending{cluster,service,deployment}
aws.ecs.service.deployment.running{cluster,service,deployment}
ECSDog will scrape all the ServiceEvents and put them in the DataDog stream so you can monitor and alert on ECS issues like scheduling problems and unhealthy instances.
ECSDog will make a best effort to de-dup events with the same ID, but you may get duplicates across daemon restarts.