Measure couch token request overhead and compare with simple caching alternative

It is not clear whether querying couch on every client request is a performance problem. We need to measure the performance of the current solution against a simple caching alternative to have a clear idea if improvement is needed.

Even if it needed, it is not clear which direction to go for caching. Twisted sessions send the session id (an no session data) in cookie request headers back and forth to retrieve session. We already have that implemented if it is to be considered as a valid option.

(from redmine: created on 2016-12-29)

• Relations:
  • relates #8667 (closed)