We are working on Tracealyzer for VxWorks, intended for release during spring 2013. This version uses the existing recorder in VxWorks (wvLib), which generates .wvr files. Thus, no target-side integration is required.
Here is a screenshot, showing a recording from quad-core system. The background shows scheduling and system calls on the four cores, and the foreground shows CPU load on Core 0 and Core 1, and a response-time plot.
(click image to enlarge)