An exceptional data query conflicted with a routine system update (index removal), causing a table lock that stopped users from being able to log into the Fleet Management Portal.
Killing the process allowed all the held off (dashboard) processes to run. This took a little while to pin it down unfortunately; analysing the client connections on the db for rogue processes, we could see that a lot of processes were held waiting for a metalock on that table.
A new automated process that monitors for exceptions queries has been introduced so this cannot reoccur.