Back in December, there were a couple of posts about API cloud rate limits. Have any of these been implemented yet, and if so, what are they? If not, what’s a “polite” rate limit to keep myself to? I would prefer not to break anything or annoy anyone
The limit for Spark.publish() is one per second average with a burst of up to four allowed. I don’t think there is a limit for functions and variables but my experience is that you can’t really call them faster than about once per second without missing something.
See my servo tutorial for an example:
If you’d like to bypass these ‘restrictions’, you could consider running a local cloud. You’ve then got full control, and can set you limits as desired. Also, it’s a lot faster, so the rates can probably go up even more.
If you do a forum search on “local cloud” you’ll be sure to find a couple of topics.
Thanks for asking, thats a great question!
In general, we want to make sure there will always be enough resources for everybody, so if you wanted to find a happy medium I would say try to only use what you need. In general, we try to only set protections when we need to protect the community or the cloud. That’s why we haven’t limited the function / variable calls yet.
This is also a question that comes up when building smart devices for commercial use as well, and we’re working on a system to help creators optimize and understand usage. As much bandwidth as one device uses, it gets a lot bigger when you make a few thousand more…
I hope that helps!