Archive for the ‘visualizations’ Category

Memory Leak (and Growth) Flame Graphs

Memory Leak Memory Flame Graph     Your application memory usage is steadily growing, and you are racing against time to fix it. This could either be memory growth due to a misconfig, or a memory leak due to a software bug. For some applications, performance can begin to degrade as garbage collection works harder, [...]

Posted on August 16, 2013 at 8:31 am by Brendan Gregg · Permalink · Comments Closed
In: flamegraphs, memory, performance, visualizations

What the Mean Really Means

When analyzing response time, or latency, you need much more information than an average provides. The average, commonly the arithmetic mean, shows the index of central tendency. But, as I found in earlier posts, the tendency is often not central, but may be skewed by outliers, or split by multiple modes. How often these factors [...]

Posted on July 12, 2013 at 10:07 am by Brendan Gregg · Permalink · Comments Closed
In: averages, frequencytrail, performance, statistics, visualizations

Modes and Modality

It is a truth universally acknowledged that the average is the index of central tendency. But what if the tendency isn’t central? I’ve worked many performance issues where the latency or response time was multimodal, and higher-latency modes turned out to be the cause of the problem. Their existence isn’t shown by the average – [...]

Posted on July 8, 2013 at 12:59 pm by Brendan Gregg · Permalink · Comments Closed
In: frequencytrail, performance, statistics, visualizations

Detecting Outliers

In computer performance, we’re especially concerned about latency outliers: very slow database queries, application requests, disk I/O, etc. The term “outlier” is subjective: there is no rigid mathematical definition. From [Grubbs 69]: An outlying observation, or “outlier,” is one that appears to deviate markedly from other members of the sample in which it occurs. Outliers [...]

Posted on July 1, 2013 at 10:43 am by Brendan Gregg · Permalink · Comments Closed
In: frequencytrail, latency, outliers, performance, statistics, visualizations

Frequency Trails

Frequency trails are a simple and intuitive visualization of the distribution of sampled data. I developed them to study the finer details of hundreds of measured latency distributions from production servers, and in particular, to study latency outliers. Latency outliers are occurrences of unusually high latency, and can be a source of performance problems. The [...]

Posted on June 19, 2013 at 6:22 am by Brendan Gregg · Permalink · Comments Closed
In: Visualization, visualizations

Revealing Hidden Latency Patterns

Latency Heat Map Response time – or latency – is crucial to understand in detail, but many of the common presentations of this data hide important details and patterns. Latency heat maps are an effective way to reveal these. I often use tools that provide heat maps directly, but sometimes I have separate trace output [...]

Posted on May 19, 2013 at 2:56 pm by Brendan Gregg · Permalink · Comments Closed
In: heatmaps, latency, performance, visualizations

USENIX LISA 2010: Visualizations for Performance Analysis

My USENIX LISA talk from 2010 is now available on youtube, also embedded below. The title is Visualizations for Performance Analysis (and more), and showed how the full distribution of data could be presented as a heat map. This is especially useful for latency analysis, as fast-path and slow-path differences can be studied, as well [...]

Posted on December 10, 2012 at 10:36 am by Brendan Gregg · Permalink · Comments Closed
In: performance, slides, video, visualizations

dtrace.conf 2012 videos

Last month was dtrace.conf 2012, the 2nd DTrace unconference. This is a meetup of DTrace practitioners and developers, where we discuss the latest uses, developments and future directions of the technology. It was great to see old friends of DTrace, and to put new faces to names. See the video list for the sessions, which, [...]

Posted on May 8, 2012 at 4:06 pm by Brendan Gregg · Permalink · Comments Closed
In: DTrace, video, visualizations

Subsecond Offset Heat Maps

“Wow, that’s weird!”. My subsecond offset visualization type looked great, but others found it weird and unfamiliar. I developed it for inclusion in Joyent’s Cloud Analytics tool for the purposes of workload characterization. Given that it was so unfamiliar, I had some explaining to do. Voxer, a company that makes a walkie-talkie application for smart [...]

Posted on March 26, 2012 at 11:16 am by Brendan Gregg · Permalink · Comments Closed
In: cloud analytics, DTrace, performance, subsecond, visualizations

Linux Kernel Performance: Flame Graphs

To get the most out of your systems, you want detailed insight into what the operating system kernel is doing. A typical approach is to sample stack traces; however, the data collected can be time consuming to read or navigate. Flame Graphs are a new way to visualize sampled stack traces, and can be applied [...]

Posted on March 17, 2012 at 9:24 am by Brendan Gregg · Permalink · Comments Closed
In: flamegraphs, Linux, performance, perf_events, systemtap, visualizations