Hello. I have two questions, trying to understand how code is moved and executed. I'm trying to anticipate network i/o, efficiency, and cost. At what points would a docker image be cached if I'm running on kubernetes? If I use s3 storage to deploy my code on kubernetes, would any sort of local agent cache the code from s3, or would it be downloaded for each flow execution?

hi @Tristan

At what points would a docker image be cached if I'm running on kubernetes?

if I'm understanding what you mean by cached, this depends on your image pull policy on your cluster. This is something you can configure on a k8s work pool (there's a field by that name). if you mean caching during build time, that sounds like a CI + docker config thing (feel free to ask a more specific question if thats what you mean)

If I use s3 storage to deploy my code on kubernetes

you would have a

pull

step on your deployment that your worker would use to pull your code each time your deployment runs

Thanks. Yep, your first reply is what I was trying to get at.