When running behind nginx on seagrant-dev, the POST cache endpoint throws a 408 request time out behind nginx. For now we can call the webapp directly on port :9002 (or any other port it is deployed on). We need to figure out whether we want to change the logic in the service to be asynchronous (preferred) or increase the nginx timeout (not preferred).

In the first case we can use a dedicated actor pool and Futures as well as maybe cap the number of concurrent connections in the postgres driver.