Built in support for partitioning of multi-step flows instead of single steps

Expected Behavior

Built in support or tags for full flows to be partitioned.

Example: a flow F consists of 5 steps s1-> s2 -> s3 -> s4 -> s5
Support for flow F to be portioned i.e. multiple instances of flow F ( s1-> s2 -> s3 -> s4 -> s5 ) could run in parallel through portioning.

Current Behavior

No specified tags to execute entire flows in parallel, although single steps partitioning can be done, there is no such support for portioning a flow of multiple steps.

Context

How has this issue affected you? My application required a flow consisting of multiple steps to be portioned, which could not be accomplished directly
What are you trying to accomplish? A batch job where each flow, consisting of series of steps, to be partitioned using Manager-Worker pattern
What other alternatives have you considered? I tried to use create custom portioning on a flow instead of step, but that did not work, and also caused several data anomalies due to sharing of resources.

Are you aware of any workarounds? Not that i am aware of

Start by examining the existing single-step partitioning and custom partitioning behavior described in the issue, along with how a multi-step flow is executed. Define how multiple complete flow instances can run in parallel under the Manager-Worker pattern, and verify that the result avoids the reported shared-resource data anomalies.