Let's consider the following server configuration:
- worker pool size: 10
- polling interval: 1h
- number of requests received since last polling run: 1000
As of version 0.1, the server will query for pending requests. It will get all the 1000 requests and submit them to the executor service. With only 10 worker threads, this may exhaust resources on the server.
It should be possible to limit the number of pending requests to take in each polling run.
This should provide throttling/back pressure on the server side.