How does rate limiting work for a multi-tenanted solution? If I create an API key per “customer” but all of the requests are coming from a single server, will that cause rate limit problems?
I believe if one of the keys falls foul of the rate limits then the IP will be limited (which will in turn limit the other keys).
Saying that, why would a multi tenanted solution on the same server require multiple connections to the (streaming or restful API) endpoints? It’d be easier to service all your ‘tenants’ through a single service (me thinks) on the server.
You may want to check out this thread: API rate limiting and Streaming API endpoints - clarification
In general if you’re using multiple keys on a single server and their cumulative requests exceed the rate limit you’re at risk of being banned at the IP level. As such you may need to build some form of internal rate limiting (or queueing). That said, we have a system like this in place, and despite having a large and active user base it’s never actually been triggered. I think most use cases where this was going to be an issue you’d probably be better off using the bulk data-sets.