A typical use case for using a shaper. . .
Suppose you have an ISP which provides a physical hand-off with much more bandwidth than you're allowed to use. Common cases would be DSL, Fios, Cable, where you might only have a faction of the physical bandwidth as an allowance. For example, you have a gig connection, but your bandwidth allowance is 200 Mbps.
Now what happens if you send 300 Mbps, on a gig connection, where your bandwidth allowance is 200 Mbps?
Generally, your ISP will drop 100 Mbps of your 300 Mbps. But do you know when this happens?
What you might do, is place a 200 Mbps shaper on the gig interface. Then you will see drops. Further, you have the option to use QoS to determine what traffic will be dropped first.
As for common example of RSVP usage, I've never seen a real world implementation.
In theory, if your network application and network are configured to support RSVP, a possible good use case example would be you with to do a tele-meeting with some across the network.
So, what should happen, your tele-meeting network application requests the network to guarantee (reserve) the bandwidth it needs between the two tele-meeting network hosts. All in-path network devices, need to participate and agree they can, and will, guarantee the bandwidth requested. If any do not, the connection is not made. But, if it is made, assuming your tele-meeting truly only needs the bandwidth it requested, it should operate flawlessly.
You might want to read this Cisco information; take note of RSVP in the IntServ model.
You might also find this Cisco TechNote helpful, as it explains a bit about how shaping, and its alternative policing, work.
I also recommend searching the Internet and/or Cisco sites, for (much) more (available) information on RSVP and/or shaping.
You also will likely find these Cisco forums most helpful when having a specific issue, like for shaping, perhaps, for example, how best to set Bc for a particular traffic type. I.e. things not often found even with extensive Internet searches.
