If you get an STP loop with a lot of traffic you will detect it very quickly because your network won't work :-)
If you do end up with a loop and it does not bring your network grinding to a halt a sure sign of it are messages in the logs about mac addresses flapping between ports although those messages do not always mean that.
In terms of preventative measures you should run BPDUGuard on your end user ports which should have portfast enabled. This means if a hub is connected to two switchports the switch will see BPDUs because the hub simply relays them and then error disable the port.
You can use port security to only allow a certain number of mac addresses per port which also means users would realise attaching a hub did not work.
In the end there is only so much you can do technically. Sometimes it comes down to user education and being able to meet their demands.
For example users generally don't connect up hubs unless they need extra ports. In one of the larger companies I worked for we had a very strict policy of no users were allowed to connect anything to the network unless authorised and it was a disciplinary matter if they did.
However we also recognised that they don't generally do it for fun so we always made sure we had spare switches in stock in case they did have a genuine need for extra ports.
I appreciate not all companies can do this but if you don't provide what they need for their work they generally will try and find a way around it.
Edit - also have a look at this link which contains some useful information for STP troubleshooting -
https://supportforums.cisco.com/document/54376/spanning-tree-loop-troubleshooting-and-safeguards
Jon
