Beta

Discover and try out LTTng projects under development.

LTTng Scope

LTTng Scope 0.4

LTTng Scope is an experimental graphical viewer for LTTng kernel and user space traces for Linux, Windows, and macOS. With this new project, we want to experiment with UI/UX and determine how users can efficiently interact with traces using a graphical user interface.

LTTng Scope 0.4 main window

Have a look at our LTTng Scope 0.3: A basic tutorial blog post for a quickstart with LTTng Scope.

Please note that LTTng Scope 0.4 is definitely not ready for use in production. We are making it available to get it "out there" and to gather valuable user feedback.

Some currently available features include:

  • Project-wide event highlighting (filter) shown in all views.
  • Thread and CPU execution state visualization.
  • Detailed event record list.
  • Timestamp formatting, interoperable with Babeltrace.
  • Streamlined workflow.
  • Convenient installer.

There are some obvious limitations during the initial development phase of this project:

  • Hard-coded set of views specific to LTTng kernel traces.
  • You can only open one trace.

We are planning to release new preview versions over the next few months as we introduce changes, new features, and performance improvements.

Do not hesitate to report any issue you encounter and give us feedback!

 lttng/lttng-scope
 Blog post: LTTng Scope 0.3: A basic tutorial
 LTTng Scope 0.4 installer for Linux
 LTTng Scope 0.4 installer for Windows
 LTTng Scope 0.4 installer for macOS

LTTng analyses

LTTng analyses 0.6.1

The LTTng analyses are a set of various executable analyses used to extract and visualize monitoring data and metrics from LTTng kernel traces on the command line.

LTTng analyses 0.6.1 frequency analysis

As opposed to other "live" diagnostic or monitoring solutions, this approach is based on the following workflow:

  1. Record your system's activity with LTTng.
  2. Do whatever it takes for your problem to occur.
  3. Diagnose your problem's cause offline (when tracing is stopped).

This solution allows you to target problems that are hard to find and to "dig" until the root cause is found.

Notable features:

  • Easy to use command-line analysis scripts with sane defaults and many configuration options.
  • Latency and frequency distribution, statistics, and full logs of various system metrics like CPU, I/O, interrupts, scheduling, and more.
  • Many filtering options: CPU number, beginning/end times, IRQ number, process name, thread ID, and more.
  • Powerful period engine to isolate the available analysis results within time intervals provided by custom conditions.
  • Human-readable output as well as JSON machine interface to postprocess the results.

Feel free to report issues you find and to give us feedback!

 lttng/lttng-analyses
 Blog post: Full Stack System Call Latency Profiling
 Blog post: Finding the Root Cause of a Web Request Latency