Skip to main content
June 16, 2021
Question

Parallel process and execution limit of 50 nodes

  • June 16, 2021
  • 25 replies
  • 0 views

Hi,

I would have a theorical question :

Let's say I have a very complex process that count 60 nodes, and the latter can not reach the end because of the 50 nodes limit.

Let's admit that 30 first nodes car be executed in parallel with the 30 remaing nodes. Does the limit of 50 nodes will occur nevertheless? 

(using an AND gateway for the 2 paths)

25 replies

danny.verb
June 16, 2021

Yes, that limit still occurs. However, using a sub-process or the start process smart service you can execute those 30 nodes in parallel and get around the limit.

June 16, 2021

Thank a lot for your reply Danny.

You mean one of these patterns ? (is it a second AND to reconcile the 2 paths?)

mikes0011
Brainy
June 16, 2021

It seems like you might be forgetting that this "limit" almost exclusively applies to continuous Activity Chaining between two user input tasks.  If user input tasks do not come into play, then the 50-node limit should almost never be relevant.

The Expression Guru
June 16, 2021

In my example, the nodes are chained together and there is no input task between the nodes.

But what do you mean ?

mikes0011
Brainy
June 16, 2021

Why do the nodes in your example need Activity Chaining enabled, then?

The Expression Guru
stewart.burchell
February 7, 2022

The guardrails are pretty clear: 50 chained nodes will be your limit, after which the synchronous experience your User was expecting will come to an end and the next thing they'll see is either:

  • an asynchronous Task in their Task list (if a Task is assigned to them after the chain limit is exceeded)
  • or the data they see in the interface from which they launched the process isn't reflecting the changes that the process actually conducted (until they refresh that interface after a point in time at which the process comes to an end)
  • or they'll not see anything that they didn't expect (in which case the process is affecting their UX or the data they're seeing, which then begs the question as to why it was chained from the 50th node onwards in the first place)

In general any process model that reaches this size probably needs to be reviewed. You can achieve a lot in a combination of Expressions, SAIL interfaces, asynchronous patterns, and even "merging" nodes (e.g. by fetching data in the 'Input' tab and then post-processing it in the ';Output' tab before passing the result to a process variable, which means many opportunities to reduce the size to less than 50 nodes.

February 7, 2022

Thank you Stewart for this further information.