Not by any reliable means...
The most reliable way to get the uptime of a host in general is to query that host... There isn't anything reliable in any given packet or packet type to indicate a host's uptime, and to further complicate that, even if you query a host, in the case of a VIP front end, are you getting the VIP's uptime or the uptime of one of the back end hosts? Which back end host?
Unless your application is specifically sending out its uptime (as one of the header records, for example), there would still be nothing to reliably say if the application is running continuously or not. A momentary break in Application Availability might be a network hiccup, an application restart, or any number of other things not actually related to the application itself, depending on where DCRUM is monitoring from. From the wire, an application restart is generally not something you'd announce (or want to).
The only thing that _could_ be semi-reliably used would be TCP connection refusals, *if* you are monitoring directly the machine the application is hosted on rather than a VIP or some other front end device. A TCP connection refusal _most commonly_ indicates that the host itself is up and running, but the application is not, though it can also mean that the application has reached its connection limit and is refusing new connections until some existing connections are closed.