Skip to main content

Has anyone by chance done a side-by-side benchmark comparison of using sub zaps vs. HTTP request & Webhook method for invoking modular workflow?  My understanding is that there is a performance hit for doing it via the HTTP request method but has anyone quantified that? Are we talking 5%...200%? 

I can see some advantages for sticking with the HTTP request / Webhook as the interface and actually wrote up a tutorial over the weekend demonstrating a useful development tactic I call “chokepoints” using Postman/Pipedream under this method. I’m curious though how big (if any) the performance penalty would be for using HTTP req vs. sub zaps at scale. 

Are there any other heuristics for when the sub zaps approach would be more favorable other than perhaps readability in the editor for less-technical folks? I saw the limitations section of this KB article indicating that sub zaps must process one and only one return so that would seem to constrain what you can do within a sub zap. TIA 

Hi @grid7 

Can you further define how you are trying to measure “performance hit”? (Best to outline a specific example.)

 

With webhooks you can create these Zap relationships:

  • 1:1
  • 1:MANY
  • MANY:1

 

I personally tend to use webhooks instead of Sub-Zaps.


Reply