Is it possible to resume Interrupts from parallel nodes

[MihaiAlx]

Hi, I'm trying to implement HITL and I have a few issues with parallel nodes.

Let's assume both parallel nodes need to ask for human input - the time it takes for each node to finish can also vary (lets say Agent 2 takes 5 more minutes than Agent 1).
Is there a way to catch and resume the interrupt for a specific node, without waiting for the other agent to finish?
Before I was using the 'human_node" aproach where every node would have a conditional edge to the human_node, it seems to not work so well with parallel nodes - are there any examples for complex HITL workflows ?

[hinthornw]

This relevant? https://langchain-ai.github.io/langgraph/how-tos/human_in_the_loop/add-human-in-the-loop/?h=resume_map#resume-multiple-interrupts-with-one-invocation

[MihaiAlx]

I might be missing something but from my current understanding - it's useful when you eventually need to resume when you have multiple interrupts in the queue (especially when all come from one node), but in the case with parallel nodes you still need to wait for the superstep to complete - kind of making the point of parallel nodes redundant