Flame graphs are a visualization of profiled software, allowing the most frequent code-paths to be identified quickly and accurately. They can be generated using my open source programs on github.com/brendangregg/FlameGraph, which create interactive SVGs. See the Updates section for other implementations.
The following pages (or posts) introduce different types of flame graphs:
The example on the right is a portion of a CPU flame graph, showing MySQL codepaths that are consuming CPU cycles, and by how much.
The x-axis shows the stack profile population, sorted alphabetically (it is not the passage of time), and the y-axis shows stack depth. Each rectangle represents a stack frame. The wider a frame is is, the more often it was present in the stacks. The top edge shows what is on-CPU, and beneath it is its ancestry. The colors are usually not significant, picked randomly to differentiate frames.
Flame graphs can be generated from any profile data that contains stack traces, including from the following profiling tools:
- Linux: perf, SystemTap, and ktap
- Solaris, illumos, FreeBSD: DTrace
- Mac OS X: DTrace and Instruments
- Windows: Xperf.exe
Once you have a profiler that can generate meaningful stacks, converting them into a flame graph is usually the easy step.
My talk for USENIX/LISA13, titled Blazing Performance with Flame Graphs, explains the visualization and summarizes different types:
These slides can also be downloaded as a PDF.
Flame graphs were released in Dec 2011. Not long afterwards (updated in 2012):
- Alan Coopersmith generated flame graphs of the X server.
- Dave Pacheco created them with node.js functions.
- Max Bruning has also shown how he used it to solve an IP scaling issue.
- Dave Pacheco has also created stackvis, a Node.js implementation, available on npm.
- Zoltan Farkas was inspired by flame graphs to create Spf4j (Simple performance framework for java), which includes them in a GUI.
More Flame Graph news (updated Apr 2013):
- I wrote some documentation for Linux Kernel Flame Graphs, generated using either perf and SystemTap for the profile data.
- Mark Probst developed Flame Graphs for Instruments on Mac OS X.
- Sam Saffron has developed Flame graphs in Ruby MiniProfiler, and shows examples of amazingly deep stacks.
- Bruce Dawson has an excellent post on Summarizing Xperf CPU Usage with Flame Graphs on Microsoft Windows. It includes examples for Visual Studio and Outlook, and a stack folding script to process the Xperf output.
- Google Chrome's performance analysis tool, WebKit Web Inspector, introduced "Flame Charts", inspired by flame graphs. These are a similar visualization, but the x-axis is time, instead of the alphabet. Check out the screenshot, which includes a mouse-over popup that links to the source code. (Update: there is a bug to add flame graphs to Chrome, in addition to flame charts: Chromium 452624.)
- Tim Bunce has been improving and adding features to Flame Graphs, and has included them in Perl's best profiler, Devel:::NYTProf, for profiling Perl. See his post on NYTProf v5 – Flaming Precision.
More Flame Graph news (updated Aug 2013):
- I wrote a document summarizing four techniques for generating Memory Leak (and Growth) Flame Graphs, which visualize stacks with byte counts, instead of the traditional CPU sample Flame Graphs. I also colored them green to indicate that they are a different type.
- John Graham-Cumming showed how CloudFlare was using SystemTap-generated flame graphs for optimizing their Lua WAF.
- Yichun Zhang showed how Off-CPU Time Flame Graphs (PDF) can solve issues of blocking time. See Off-CPU Analysis for why this is important.
- Igor Soarez wrote How To Make Flame Graphs for node.js analysis, showing all steps involved.
- Paul Irish and Umar Hansa posted an awesome demo of using FlameCharts to investigate time in V8, which includes zooming in and clicking on functions to go to code, and Addy Osmani posted a longer video tutorial. While these aren't Flame Graphs, they show developments in a related visualization: a time-series version that retains sequence and ordering.
More Flame Graph news (updated Jan 2014):
- I gave a talk at USENIX/LISA13 titled Blazing Performance with Flame Graphs, which covered them in enormous detail. See the slides and video. At the last minute this talk became a Plenary when another speaker became ill and had to cancel.
- Joab Jackson wrote an article about flame graphs for PCWorld, InfoWorld, and CIO.
- I wrote documents on Off-CPU Flame Graphs, and Hot/Cold Flame Graphs, based on what I had documented earlier for my USENIX talk.
More Flame Graph news (updated Jun 2014):
- I wrote a post showing how the Google lightweight-java-profiler can be used to create Java Flame Graphs. This is only Java code time, and not other CPU consumers. See my later work with Java (-XX:+PreserveFramePointer) for a different and more complete profile.
- Vladimir Kirillov has developed eflame (github), a Flame Graph profiler for Erlang. Blocking calls are in blue, see the screenshot.
- Trevor Norris posted instructions for creating Linux Node.js CPU flame graphs, by using perf_events and --perf-basic-prof. See the example.
- Luca Canali provides many examples of Flame Graphs for Oracle databases, and explains what they mean.
- Julien Charbon posted some CPU flame graphs on FreeBSD, to analyze TCP stack lock contention with short-lived connections.
More Flame Graph news (updated Nov 2014):
- Facebook's Strobelight uses an inverted flame graph for perf analysis. See the slide in Yannick Brosseau's Using tracing at Facebook scale (PDF) talk.
- The post on Profiling a Meteor app: Telescope includes flame graphs.
- Adrien Mahieux added the Click to Zoom feature to the original FlameGraph software. This is really awesome!
- I wrote a post on Differential Flame Graphs, showing new features I added to flame graphs for performance regression testing.
- I came up with an unusual but useful use for differential flame graphs: CPI Flame Graphs, to highlight memory stall cycles.
- I gave a talk at the 2014 FreeBSD Developer and Vendor Summit on Flame Graphs on FreeBSD.
- I wrote a post on Node.js Flame Graphs on Linux using Linux perf_events and v8's --perf-basic-prof option.
- Yunong Xiao showed how flame graphs helped solve an important production performance issue for Netflix in the post Node.js in Flames.
- Oozou developed and published RubyProf::FlameGraphPrinter for ruby-prof, to generate folded stacks for flame graph generation.
- Thorsten Lorenz developed a web flamegraph interface that can load new profiles, and perform a regexp search. See his example.
More Flame Graph news (updated May 2015):
- Shawn Sterling posted about his excellent talk at Linuxfest Northwest 2014: Getting Started With Flamegraph, which includes both the slides and video (turn the volume up).
- Scott Lystig Fritchie gave a great talk on profiling Erlang code including flame graphs at Erlang Factory San Francisco 2015.
- I submitted tickets for a flame graph / flame chart toggle for Google Chrome (Issue 452624) and Firefox (bug 1123495). I'd love to see both implemented. The chromium ticket prompted interesting discussion and prototypes.
- In my SCALE13x talk (2015), I previewed mixed-mode Java flame graphs using Linux perf_events ("perf"). This uses a JVM frame pointer patch that has become the -XX:+PreserveFramePointer option in both JDK9 (JDK-8068945) and JDK8u60 (JDK-8072465). For the first time, we can see all CPU consumers in one visualization! See slides 40 to 57 for more details, and my post about it (when I write it!).
- Francesco Mazzoli post about Flame graphs for GHC time profiles with ghc-prof-flamegraph, including a tutorial for making flame graphs from the existing GHC (Haskell) profiler output.
- Robin Moffatt posted about Analysing ODI performance with Flame Graphs, for understanding load plans in the Oracle Data Integrator.
- Flame graphs were demoed at d3NYC (Drupal) 2015, but I haven't found slides online yet.
- Cor-Paul Bezemer has been investigating flame graph differentials with his Flamegraphdiff software, which shows the difference from A to B using three flame graphs simultaneously. This was also the subject of a SANER2015 paper ($), and talk.
- M. Isuru Tharanga Chrishantha Perera wrote a translator for Java Flight Recorder profiles, for making flame graphs.
- Jan Stępień developed Pluggable Flame Graphs for Clojure (example).
More Flame Graph news (updated Jul 2015):
- Evan Hempel has created a python-flamegraph profiler for Python, which generates the folded stack output suitable for making into flame graphs. See the README on github for instructions.
- Franck Pachot posted another example of using CPU flame graphs for identification of an Oracle database bug.
- Strongloop have included a nice looking Node.js flame graphs module for their Arc Profiler.
- @yoheia posted a detailed flame graph post showing Linux kernel profiling (in Japanese).
- Zoltan Majo fixed JDK-8068945 for Java 9 and JDK-8072465 for Java 8 update 60 build 19 (or later, download as early access here). This adds -XX:+PreserveFramePointer, which allows Linux perf_events to sample full stacks for making flame graphs. This began with a prototype patch I developed and submitted: (A hotspot patch for stack profiling). Great to see this functionality make it into the JVM!
- Min Zhou has developed PerfJ, for automating the collection of Java flame graphs, using the frame pointer patch (see previous item).
- Jonathan Perkin used memory flame graphs for Reducing RAM usage in pkgin.
- Isuru Perera postabout about Java CPU Flame Graphs with an example making use of Min's PerfJ.
- Ben Sandler from Uber has posted go-torch, a flame graph profiler for go-lang programs.
- Myself and Martin Spier posted about Java in Flames (PDF) for the Netflix Tech Blog, showing Java mixed-mode flame graphs using the new -XX:+PreserveFramePointer JVM option.
More Flame Graph news (updated Aug 2015):
- Andi Kleen demonstrated Generating flame graphs with Processor Trace, a feature from modern Intel CPUs for very high frequency sampling.
- I wrote about the new flame graph search feature. The matched percentage is very handy, so I don't have to mouse over many tiny frames and add them up manually.
- Eben Freeman posted about Profiling Python in Production, which includes a python profiler and basic d3 flame graphs.
More Flame Graph news (updated Oct 2015):
- Will Sewell mentioned flame graphs in his Top tips and tools for optimising Haskell post, and linked to the work for GHC flame graphs by Francesco Mazzoli.
- Bo Lopker created djdt-flamegraph (github) for getting a flame graph of current requests in the Django Python web framework.
- NodeSource have flamegraphs in their NSolid Node.js analysis product. The graphics look very nice, and they also have treemaps and sunbursts (both of which I think are usually less effective than flame graphs, but I don't mind having the option).
- Alex Ciminian has been developing a d3-flame-graphs library (github) for d3, implemented in CoffeeScript. Check out the demo, it looks really nice so far.
More Flame Graph news (updated Dec 2015):
- I gave a Java Mixed-Mode Flame Graphs talk at JavaOne 2015, including all the latest updates.
- David Calavera wrote about Docker flame graphs for Go.
- Carol Nichols wrote detailed instructions for Rust Profiling with Instruments and FlameGraph on OSX: CPU/Time.
More Flame Graph news (updated Apr 2016):
- Apekshit Sharma wrote a post about Saving CPU! Using Native Hadoop Libraries for CRC computation in HBase.
- Mike Huang wrote a Netflx tech blog post to show how he helped with Saving 13 Million Computational Minutes per Day with Flame Graphs.
- Srdjan Marinovic and Ryan Day created goprofui includes a cpu.go profiler for golang and flame graphs.
- I wrote an article for ACMQ, The Flame Graph, which defines flame graphs, describes their origin, explains how to interpret them, and discusses possible future developments.
More Flame Graph news (updated Jun 2016):
- David Mark Clements has developed 0x, a new interactive flame graph profiler for Node.js processes on both Linux and OS X. See the demo.
- Qt Creator 4.0.0 now includes flame graphs along with timeline and statistical views. That's how it should be done: different views of the same data you can switch between.
- Antonio Pérez developed 4gl-flamegraph for processing the profiler output of Genero by 4Js (Four Js).
- Ty Overby published a flamegraph profiling tool for rust.
- The vprof Python package provides interactive visualizations for profiling, including flame graphs.
- Alastair Butler and Kei Yoshimoto published the paper Large scale semantic representation with flame graphs (PDF), where they used them to visualize semantically annotated corpora, which has been parsed into a hierarchy.
- Jonathan Newbrough developed the Gumshoe Load Investigator for Java analysis, initially for internal use in the Dell Cloud Manager, which makes good use of flame graphs.
- Mahesh Dathrika published Igniting Node.js Flame, showing how they use the v8-profiler at eBPF to create Node.js flame graphs.
- Maciek Lesiczka is building an ASP.NET profiler that includes flame graphs, and introduced it in netric.io - lightweight ASP.NET profiler.
- Nitsan Wakart covered Java flame graphs and various Java profiling challenges in his Devoxx UK talk Extreme Profiling: Digging Into Hotspots by Nitsan Wakart (youtube).
More Flame Graph news (updated Jul 2016):
- Maciek Lesiczka has been developing netric.io - lightweight ASP.NET profiler, an ASP.NET profiler that includes flame graphs.
- Rhys Hiltner from Twitch blogged about Go's march to low-latency GC, which included flame graph analysis of GC time.
- Dmytro Semenov on the ebay tech blog posted Mastering the Fire, about how they can create flame graphs on production servers at any time with one click of a button.
- Olivier Cano posted about Measuring Snap performance which includes flame graphs for golang.
- Kay Ousterhout wrote about Generating Flame Graphs for Apache Spark using Java Flight Recorder.
- Adam Perry published a post about Rust Performance: A story featuring perf and flamegraph on Linux, which also has great examples of using perf.
Thanks to everyone who has written about flame graphs, developed them further, and shared their results! I'll update this page from time to time with more news.