The trace mechanism provided and used by MQe has the following features:
- A pluggable interface to allow user-written trace handlers to be implemented
if required.
- A variety of implementations of the trace handler interface to cater for
a variety of uses. One such implementation supports a crude form of circular
logging, so older trace information is discarded when newer trace information
becomes available. See the com.ibm.mqe.trace.MQeTraceToBinaryFile for
more details.
- A separation between the trace point number, and the meaning, or textual
representation of that trace point. This separation of the number from lengthy
meaningful string information allows for collection of the trace point numbers
to be performed at runtime, and the rendering of that information to a readable
format to be done offline at a later time. This can mean trace information
files are smaller and generated more quickly at the point of capture, but
much larger and more accessible at the time they are read.
- Dynamic, runtime filtering of trace information.