track browser and tab for staff interface logging

[Galen Charlton]

It would be handy if the Evergreen staff interface could uniquely identify the browser and tab that originated a request. Doing this would allow:

- easier filtering of log entries from a specific browser when debugging issues
- providing a hook for server-side rate-limiting if a specific browser is spamming too many requests (as has been a persistent problem)

A browser ID could be a (short!) UUID generated and stored in local storage or a cookie, while a tab ID could be maintained in session storage.

Opening this as both an Evergreen and OpenSRF bug since logging such identifiers would entail either adding new optional attributes or elements to the OpenSRF message or including them as part of the message XID (which probably implies making some changes to the OpenSRF JavaScript client libraries to ease management of the XIDs if the staff client becomes responsible for always providing them.)

Note that a mechanism like this is proposed for the staff interface only (and maybe as an option for third-party API clients), as storing and sending quasi-persistent browser and tab IDs for OPAC requests would present patron privacy issues.