Michael Hart

(that's why you see all the pings btw – because it takes so long to init – they ping every 100ms so usually you don't see them)

Haha, lol – yeah, that'd be it – didn't even think of it

Oh, we can't just do it behind a flag? Like, either `import()` or `require()` depending on an env var?

@brandonryan is the loader stuff only an issue if ESM is used? If so, we could just leave the experimental stuff up to the NODE_OPTIONS env var if ppl wanted...

I feel like the best way to provide backwards compat is to default everything here to be in CommonJS – ie, bootstrap.js uses requires, etc. Then `NODE_PATH` will work as...

Yeah, absolutely agree – once the dust settles, will look more into this 👍

@faceleg is one of your builds still running on that machine? It could be this command that's failing: https://github.com/lambci/ecs/blob/master/runbuild.sh#L20 (basically, the `--rm` flag is trying to remove any existing containers...

Do you know what version of the ECS agent is running on your machines? I know that newer agents will try to clean up old images, etc: http://docs.aws.amazon.com/AmazonECS/latest/developerguide/automated_image_cleanup.html Might just...

Do you know what version of the ECS agent is running on your machines? Have you tried with one of the newer AMIs? http://docs.aws.amazon.com/AmazonECS/latest/developerguide/ecs-optimized_AMI.html

@RedaBenh bit strange – doesn't really tell you what the problem was aside from the fact that the EC2 instance didn't start up within 5 minutes. Maybe just try again...