Ralf Santa Cruz
10/26/2023, 9:01 PMKevin Grismore
11/03/2023, 2:32 PMrun_deployment
, which you can use in either a blocking fashion (wait for this deployment to finish, then call it again with different params to run the next deployment) or to run many child deployments in parallel, or a mix of both.
Sometimes having an entire separate pod for each unit of work can feel like overkill, so my only additional recommendation is to verify there's a good justification for doing so, like dependency conflicts that necessitate different images, or deadline constraints on compute-heavy workloads that necessitate parallel runs.Kevin Grismore
11/03/2023, 2:33 PM