I'm trying to use

execute_pipeline_iterator

or

execute_job

to launch a dagster run from within an op. I do this by passing the existing

instance

which I get from the

OpExecutionContext

of the parent op, this works and I can see the new run happening in the dagster UI, however i want to be able to control the execution of the child job, i.e. in a k8s context i want this to happen on a new pod, however by default this is happening on the same pod as the parent op. Is there something I can do here, or is there a different approach I should be taking?

hi @Clement Emmanuel! the

execute_job

python api is fairly literal -- it executes the job within the current process. Instead, you could consider using the dagster graphql python client. This allows you to submit a job to be executed, and is functionally identical to hitting the "launch run" button in the UI (and so will have identical execution behavior)

Thanks! That's actually exactly what I realized last night and it's working for me, good to hear that as validation from you as well