Often we land into issues and Support engineers respond to tickets asking for logs and forward to engineering or PG groups, why not collect all that you can.. here is some strategy I propose, come up with the integration of error and data capture with in portal - upon any user impact failures, resulting in a popup like "we experienced a hiccup, do you wish to retry" with an unique reference id and checkbox which packages the session data and uploads to microsoft servers(under user name) for future reference and ticket logging.. this approach would be much intuitive and user friendly.. this will help reference the issues encountered by ID and lookup