scalopus_examples

scalopus_examples

readme_example.cpp

The most important example is the readme_example.cpp. This example shows how the C++ tracepoints should be used and how the transport and endpoints at the process under test should be setup.

This example is built for all three tracing backends. Resulting in the following binaries:

• readme_example_native
• readme_example_lttng
• readme_example_nop

random_callstacks.cpp

This example is not really an example of how to use Scalopus, it is more an executable that can create artificial callstacks at the rate and thread count specified. The maximum callstack depth will be ten.

It may accept arguments:

./random_callstacks_native [time_base (usec) [thread_count]]
Where time_base defaults to 10000 usec, resulting in scopes that have a self time of (2 + 2 + 1) * 10 ms = 50 ms.
The thread_count defaults to 1.

This example is built for all three tracing backends. Resulting in the following binaries:

• random_callstacks_native
• random_callstacks_lttng
• random_callstacks_nop

The output from this example looks like:

View locally by opening chrome://tracing and loading trace_random_callstacks.json.gz.

embedded_catapult_server.cpp

This shows how one could embed the catapult server in the process that produces the tracepoints. This example is only built using the native tracing backend, because that's the only one that makes sense for such a usecase. I don't expect this use case to be particularly useful, but it's possible.

embedded_catapult_recorder.cpp

This shows how one could embed the catapult recorder in the process that produces the tracepoints. This example is only built using the native tracing backend, because that's the only one that makes sense for such a usecase. This can be useful when measuring durations in short-lived programs, or when the full lifetime of the process needs to be captured.

query_servers.cpp

This results in a binary that will connect to all discovered servers and obtain the data provided through the introspect , process info and trace mapping endpoints. The binary outputs human readable text to stderr while it is working, at the end it outputs the data it collected as pretty printed json to stdout.

showcase_toggle_tracepoints.cpp

This showcase how one could use the TRACING_CONFIG_THREAD_STATE_RAII macro to disable tracepoints conditionally depending on the code path taken.

showcase_marker_events.cpp

This is an example of the marker events that span the global, process or thread scopes.

showcase_count_events.cpp

This allows displaying graphs at the top of the process.