How to find and select the part of the workflow, that is not busy?

Hi community,

to avoid long wait times or status errors, i would need a workflow, that chooses, which instance is not running and waits for a minute, if all are busy…

Sorry do not understand. Also will always just one node be running at a time, and it will always be the same one as always just one node after another executes and only starts the next one once the previous one finished.

Hi Jan, the nodes are just for illustration. I am not sure which ones to use.

The system should check instance 1 → 2 → 3 … and put the workload on the next instance, that is idle.

And how would that check if an instance is busy look like? Is there an HTTP endpoint that you can call and then returns if it is busy or not?

I actually don’t know. I am a real beginner with n8n. I hope for ideas.

Sorry, now I am lost again. We can for sure help on the n8n side but if we do not know how the service works that you want to connect to, there is not much that we can do

oh - xD sorry for missunderstanding. The service is openAI. The openAI nodes are the correct ones. The other nodes i just put there for illustration, since i don’t know how to realize this functionality.

Hi again @jan . Is it more clear now? Do u have an idea, what it could be? I still need a solution.

Actually if that is the case, then my original answer applies again. n8n does execute one node after another. So there is no executing 3 nodes in parallel and then choosing the one that is “less busy” or the “fastest”.

Hi again Jan, thx for your answer. In my case, this is actually almost a dealbreaker for using n8n. Do you maybe know a workaround?

I was thinking about something like this:

I’d really like to continue using n8n…

There is currently sadly nothing like that in core n8n. Was honestly also nothing anybody every asked about as much as I know. So if you would need something like that there are probably only two ways to make it work, and both would require some coding knowledge:

  • Create a custom n8n node which does that underneath the hood
    or
  • Use a Code Node and make the requests in there, and then return the data of the first request that resolves (the Code Node supports promises)

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.