Best Practice for XRay Daemon - run one in each ECS task or create one for all our components to use?

Question

I've been struggling with the concept of running one single XRay Daemon ECS container as a "hub" for all of our ECS tasks and EC2 instances  instead of running a sidecar in every ECS task.  Granted, the sidecar images are not that big, but there could end up being a lot of them.  I guess I'm just reaching out for some opinions on which method is best:
Side-car daemons in every ECS task - or - go for creating one XRay Daemon instance as a "hub" through which all our XRay tracing segments will pass.

Answer

The recommended way is to configure the X-Ray daemon using the daemon service scheduler type, as described in these docs: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ecs_services.html

This way, you don't need to bother with the costs & configuration associated with deploying the daemon as a standalone service behind a load balancer. It deploys one daemon per container instance automatically, which means all your tasks will be able to communicate with the daemon on their instance using localhost as intended."

You may also refer to this Github issue [1] for more insights.

References:
[1] - https://github.com/aws/aws-xray-daemon/issues/53

Best Practice for XRay Daemon - run one in each ECS task or create one for all our components to use?

Relevant content