Hey All, just wanted some clarification since I couldn't find any in the docs for this - what is the difference in pipelineRun section of the helm chart vs the user-code-deployment chart? I always thought the image which was hosting the grpc server would also let the daemon know which image to execute with

Yeah, if you set user-code-deployment (which is the recommended path now) you should not need to worry about pipelineRun, your understanding is correct.

thanks dan!

@daniel following up on this, does that mean we can completely comment out the pipelinerun section? since i’ve noticed if we set the repo both in usercode and runLauncher in the chart, the latter takes precedence in defining run env at least, with my evidence supporting this being the run pods don’t see k8 secrets if added to the grpc section but do see them if added to the runLauncher section. the basic need is multiple workspaces within one k8 cluster and being able to start run pods with a certain code repo image, runtime vars, and dagster execution args per a specific workspace @Diana Stan

Yeah, my recommendation is to not set the 'image' key in the runLauncher in the helm chart, for the reason that you mention. I think if you proceed as though pipelineRun didn't exist (i.e. leave it out of your values.yaml and allow it to set the default values) then things should work the way you are hoping them to.