Hi - I am new to Metaflow, had the below question. Please direct me to the right documentation/topic incase the question is already answered in same. - Thnx Is there a way to trigger (or deploy) multiple flows within same container/pod? For eg., I have a set of flows F1, F2 and F3 with each having its owns steps within. And I have a pod already running in my cluster, is there a way to execute all 3 (F1, F2 and F3) in that already running pod? Or each flow always will have to spin off its own pod/container?

Hi Matthew, if you’re using

@kubernetes

or

--with kubernetes

, each step of the flow would launch its own pod. If I may, what’s the use case you have in mind for running it inside the same (already running) pod?

Hi Utkarsh. We have few models (lets say M1, M2, M3, corresponding to flows F1, F2, F3) that does different types of predictions on the same data and has to share the input & intermediate data states between them. (eg. M2 uses user input and M1's output if available). And each of these can be triggered(called) separately as well by the client. (M2 can be called without M1 and M3). We have 20+ models used to process the same input and some are very lightweight models and doesn't need its own dedicated pods, and hence trying to see how we can optimize on pods vs communication overhead between each flow for the same input. Hope this makes sense.