We have a microservice application which interacts with a web API. That web API implements IP based throttling. In order to bypass that throttling, we have to make sure our app sends requests to that web API from multiple public IPs.
- How could one approach this problem to ensure the set of IPs used can grow with our cluster size?
- If that is not possible, let's assume I create 20 elastic IPs, how could I make sure the load of my cluster is evenly distributed across these IPs so that I can in effect send 20 times more requests to that throttling service?