flamegraph | Flamegraph profiling support for Ruby | Monitoring library

I'm trying to export update() sub-function from following module which looks like:

I would like to know that the Python interpreter is doing in my production environments.

Some time ago I wrote a simple tool called live-trace which runs a daemon thread which collects stacktraces every N milliseconds.

But signal handling in the interpreter itself has one disadvantage:

Although Python signal handlers are called asynchronously as far as the Python user is concerned, they can only occur between the “atomic” instructions of the Python interpreter. This means that signals arriving during long calculations implemented purely in C (such as regular expression matches on large bodies of text) may be delayed for an arbitrary amount of time.

Source: https://docs.python.org/2/library/signal.html

How could I work around above constraint and get a stacktrace, even if the interpreter is in some C code for several seconds?

Related: https://github.com/23andMe/djdt-flamegraph/issues/5

Current libraries for making "interactive SVGs" in the web-browser are actually using javascript to manipulate a svg object in the page dynamically but don't actually embed the full javascript needed to animate the SVG in the svg tag itself. Libraries like d3.js, protovis, svg.js, etc.

But it's possible to create standalone interactive SVGs, for example Brendan Gregg's flamegraph tool, example: http://www.brendangregg.com/FlameGraphs/cpu-bash-flamegraph.svg

Is there any way to create a standalone, interactive SVG in a webpage? Of course, you could generate the interactive SVG serverside (for example, Brendan uses a Perl library to generate the interactive SVG) and then serve that to the user. But I'd prefer to generate the SVG as part of generating the webpage itself, i.e. in Php or Ruby or, even better, client-side with a Javascript library. In either case, the key feature goal is to be able to right-click-download the custom-made interactive SVG.

Note: The benefit of the d3.js approach is that the animation can be integrated between the SVG and other DOM elements on the page, or have interaction with the SVG trigger dynamic AJAX requests. I'm not expecting the stand-alone SVG to be able to do these things.

Google Chrome and also the NodeJs inspector allow to generate a cpu profile with the following JSON structure:

excerpt

I am using Rails and am experiencing a connection pool error very randomly and it does not target any single endpoint specifically. I can hit endpoints about 70% of the time without getting this error. The database is PostgreSQL running on Google Cloud. Here's the main stuff of the error I'm getting:

Currently, I can tell trace-cmd (an ftrace front-end) to use ftrace's function graph infrastructure to trace a specified workload. Using some trace data processing tools, I can look at the frequency with which functions were called and the corresponding total duration across each of the function calls. Example output might look like the following, where I trace ext4 function calls generated by my workload:

I am doing Profiling on my NodeJs app, I am using Google App Engine Flexible and for it, I am using npm 0x, but the thing is as this package is making the flamegraph inside my NodeJs root directory folder, now how can I retrieve this folder(I mean get access to this folder). I have SSH to my App Engine Flexible instance but there were two folders vm-runtime-app, vmagent but there my NodeJs Source code root directory is not there.

UPDATE 20180209: I created a new dummy app with what comes with the initial Rails 5.1.4 install (which includes minitest 5.11.3) and tests completed without issue. Going to continue to experiment with the gemfile to see if I can't narrow down what in the world may be causing this problem...

If you have any ideas or pointers, please let me know!

I recently started testing a new portion of my codebase and I've discovered a very critical issue. For whatever reason, every test is now throwing the titled error. Looking into the trace, it looks like for some reason the quoted name/title of my tests is causing a problem. Going back and running tests that I know were passing (though admittedly on earlier gem versions) unfortunately yields the same results, as shown below with the User model.

I've tried everything I know to do, including completely removing and reinstalling RVM, Rails, and Ruby. I've even tried versioning Minitest to an earlier build, like 10.5.3, but still no luck. I've spent most of the day trying to fix this; I'm at my wits end and desperately need help! Though I've learned the hard way that I now need to explicitly version protect ALL of my gems ;)

Thank you in advance! Please let me know if you need me to include any more files/snippets. I'll do my best to reply to any questions as my schedule allows, hopefully within 24 hours.

Returned trace:

E

Error: UserTest#test_should_be_valid: TypeError: no implicit conversion of nil into String

/home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/railties-5.1.4/lib/rails/test_unit/reporter.rb:70:in method': undefined methodtest_should_be_valid' for class Minitest::Result' (NameError) from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/railties-5.1.4/lib/rails/test_unit/reporter.rb:70:informat_rerun_snippet' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/railties-5.1.4/lib/rails/test_unit/reporter.rb:23:in record' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:803:inblock in record' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:802:in each' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:802:inrecord' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:334:in run_one_method' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:321:inblock (2 levels) in run' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:320:in each' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:320:inblock in run' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:360:in on_signal' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:347:inwith_info_handler' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:319:in run' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/railties-5.1.4/lib/rails/test_unit/line_filtering.rb:9:inrun' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:159:in block in __run' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:159:inmap' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:159:in __run' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:136:inrun' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/minitest-5.11.3/lib/minitest.rb:63:in block in autorun' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/spring-2.0.2/lib/spring/application.rb:171:infork' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/spring-2.0.2/lib/spring/application.rb:171:in serve' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/spring-2.0.2/lib/spring/application.rb:141:inblock in run' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/spring-2.0.2/lib/spring/application.rb:135:in loop' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/spring-2.0.2/lib/spring/application.rb:135:inrun' from /home/blake/.rvm/gems/ruby-2.4.1@land_app/gems/spring-2.0.2/lib/spring/application/boot.rb:19:in ' from /home/blake/.rvm/rubies/ruby-2.4.1/lib/ruby/site_ruby/2.4.0/rubygems/core_ext/kernel_require.rb:55:inrequire' from /home/blake/.rvm/rubies/ruby-2.4.1/lib/ruby/site_ruby/2.4.0/rubygems/core_ext/kernel_require.rb:55:in require' from -e:1:in'

This is from a fairly standard User model test that has not raised issues before...

user.rb

Firefox Developer Tools support flame charts, but it seems time is used as X-axis:

In the Flame Chart view itself, along the X-axis is time. The screenshot above covers the period from 1435ms to a little past 1465ms. Along the Y-axis are the functions on the call stack at that point in time, with the top-level at the top, and the leaf function at the bottom. Functions are color-coded to make them easier to distinguish.

Usually, flame charts use percentage as X axis, for instance: pyflame or Brendan Gregg's flame graphs.

I find the time-based flame charts confusing and miss the ability to get aggregated times per function (similar to Call Tree as far as content is concerned, but represented in more visual way). Also, if a function is short but is called many times it is easy to miss it, while on "aggregated" flame charts it would stand out prominently. To see why time-based flame charts are confusing, let me borrow example from another post - admittedly in Chrome, but the problem is similar:

Even if some function took 30% of total time, that would not be visible in the above chart.

Is there any way (through plugin or setting) I can make Firefox display aggregated flame charts for the selected time slice instead of time-based ones?