We will continue to spawn up a benthos pod for every data flow component or protocol converter. the resource overhead of benthos is quite small, and its better for reliability so if one flow has issues (e.g., is taking to much CPU) we can limit its impact on that single flow (e.g., resource limit on a pod) and it does not take down all the other flows. I think this is required in a multi-user scenario, that a bad pod from a single user cannot take all other pods down (on purpose, or by accident) the feature you described is what we are planning for "templating" in the management console.