The graphs on TL-SG2008 show bandwidth utilization absolutely correct. You won't see much bandwidth utilization if surfing or streaming music, of course. You even won't see much bandwidth utilization if you would have a 1 Gbps connection to the Internet when streaming music, just b/c streaming music doesn't consume significant bandwidth (not to speak of browsing the web, which also requires very low bandwidth, despite the fact that WWW sometimes seems to stand for World Wide Waiting with some bloated websites nowadays).

If I watch a movie served from a NAS in my LAN, it even shows very low utilization of the switch (and I very much appreciate low utilization for streaming movies in a Gigabit network, otherwise there would be some serious fault in the network config ;) ).





If I copy the same movie instead of streaming it, bandwidth utilization increases, but still doesn't saturate the network (again, I really don't want it to saturate the network). For example, this is bandwidth utilization of the same movie as above, but now copied at once rather than being streamed in real-time:





If you do real performance tests - e.g. with iperf -, you will get graphs like this (where 620 Mbit/s was the practical limit for transferring data between the specific systems used for the tests, which were under load caused by other applications when testing):





So there is absolutely no reason to blame the switch if it shows you an actual bandwidth in the 0 to 1% range while streaming music or movies over the Internet. It is just so, b/c streaming doesn't need much bandwidth. If you want to measure relative bandwidth usage in respect to the most achieved in a certain time range, which then is considered 100% of utilization, you need another statistics utility such as RRD tools.

BTW: 200 GB per month is about 647 Kbit/s on average.