Full trace

The full trace files are written as standard text files and should be readable in all common text editors.

The full trace output is more like a traditional debug trace. However, the detail is limited to the granularity of the instrumentation API. This makes it detailed enough to diagnose many problems while still being reasonably easy to follow.

The format of each line in the file is:

<Thread-Id>  <UnitOfWork-Id>  <time-nanoseconds>  <Activity-Type> : <info>

where:

<Thread-Id>

is the ID of the thread as reported by the JVM.

<UnitOfWork-Id>

is an arbitrary ID that is unique in the context of the execution group.

<time-nanoseconds>

is a nanoseconds value formatted as:

seconds.milli micro nano

<Activity-Type>

is a string representing the type of activity being logged.

<info>

has a format dependent on the type of activity being reported.

The file is thread safe in that each entry is issued in a synchronized manner, but the file may contain inter-thread entries. The ability to see how different threads may be interacting is important, but if you need to sanitize the file you can simply pass the file contents through a text search utility using <Thread-Id> or <UnitOfWork-Id> as a search term.

The following figure shows an extract from a trace report.

Figure 1. Example of a full trace report