The increasing complexity of automotive real-time applications, and the challenges arising with the change to multi-core processors put higher demands on the tools which are involved in the development of such systems. With tracing it is possible to record the dynamic behavior of time-critical applications and use the data to monitor safety-critical requirements. However, not all trace techniques are sufficient for this use-case. Several factors like the available hardware, bandwidth, timing accuracy, and number of traceable objects must be considered. Additionally, traces are usually not recorded on a level that is suitable for timing analysis. Hence, a transformation of the trace may be necessary. In this paper we give an overview of existing trace techniques and discuss their applicability for the timing analysis of embedded systems. We also take limitations which may be caused by existing hardware platforms into consideration. Based on a real industrial project we demonstrate how to work around common problems and show the implications of these workarounds. Next, we discuss how raw low-level trace data can be transformed to a level that is helpful for the timing analysis by understanding the internals of the operating system. Finally, we show how the resulting trace allows the validation of timing requirements and other related use-cases.