Someone wrote into support and asked the following question:
I didn’t see anything about minimum pacing with the API. Is sleeping 1/100th of a second before each parallel request of up to 30 maximum okay?
Here’s the answer:
We heavily cache all of our API endpoints, so we don’t mind if you hit them pretty hard, because we will simply return the same data until they are updated. For example, our prices update every 10 seconds, so you will only receive new data every 10 seconds. The rest of the time you’ll get the same response. However, you may find that this is wasteful of your resources because you are making far more requests that necessary, when instead you could be re-using that data. To help with this process, we include standard HTTP caching headers in all our responses. You can obey those headers and cache responses locally to optimize your side and keep your workload as light as possible.