Microservices Expo Authors: Elizabeth White, Ruxit Blog, Derek Weeks, Dana Gardner, TJ Randall

Related Topics: Java IoT, Microservices Expo, Adobe Flex, IoT User Interface, Apache

Java IoT: Article

Why Averages Are Inadequate, and Percentiles Are Great

Averages are ineffective because they are too simplistic and one-dimensional

Anyone who ever monitored or analyzed an application uses or has used averages. They are simple to understand and calculate. We tend to ignore just how wrong the picture is that averages paint of the world. To emphasis the point let me give you a real-world example outside of the performance space that I read recently in a newspaper.

The article was explaining that the average salary in a certain region in Europe was 1900 Euro's (to be clear this would be quite good in that region!). However when looking closer they found out that the majority, namely 9 out of 10 people, only earned around 1000 Euros and one would earn 10.000 (I over simplified this of course, but you get the idea). If you do the math you will see that the average of this is indeed 1900, but we can all agree that this does not represent the "average" salary as we would use the word in day to day live. So now let's apply this thinking to application performance.

The Average Response Time
The average response time is by far the most commonly used metric in application performance management. We assume that this represents a "normal" transaction, however this would only be true if the response time is always the same (all transaction run at equal speed) or the response time distribution is roughly bell curved.

A Bell curve represents the "normal" distribution of response times in which the average and the median are the same. It rarely ever occurs in real applications

In a Bell Curve the average (mean) and median are the same. In other words observed performance would represent the majority (half or more than half) of the transactions.

In reality most applications have few very heavy outliers; a statistician would say that the curve has a long tail. A long tail does not imply many slow transactions, but few that are magnitudes slower than the norm.

This is a typical Response Time Distribution with few but heavy outliers - it has a long tail. The average here is dragged to the right by the long tail.

We recognize that the average no longer represents the bulk of the transactions but can be a lot higher than the median.

You can now argue that this is not a problem as long as the average doesn't look better than the median. I would disagree, but let's look at another real-world scenario experienced by many of our customers:

This is another typical Response Time Distribution. Here we have quite a few very fast transactions that drag the average to the left of the actual median

In this case a considerable percentage of transactions are very, very fast (10-20 percent), while the bulk of transactions are several times slower. The median would still tell us the true story, but the average all of a sudden looks a lot faster than most of our transactions actually are. This is very typical in search engines or when caches are involved - some transactions are very fast, but the bulk are normal. Another reason for this scenario are failed transactions, more specifically transactions that failed fast. Many real-world applications have a failure rate of 1-10 percent (due to user errors or validation errors). These failed transactions are often magnitudes faster than the real ones and consequently distorted an average.

Of course performance analysts are not stupid and regularly try to compensate with higher frequency charts (compensating by looking at smaller aggregates visually) and by taking in minimum and maximum observed response times. However we can often only do this if we know the application very well, those unfamiliar with the application might easily misinterpret the charts. Because of the depth and type of knowledge required for this, it's difficult to communicate your analysis to other people - think how many arguments between IT teams have been caused by this. And that's before we even begin to think about communicating with business stakeholders!

A better metric by far are percentiles, because they allow us to understand the distribution. But before we look at percentiles, let's take a look a key feature in every production monitoring solution: Automatic Baselining and Alerting.

Automatic Baselining and Alerting
In real-world environments, performance gets attention when it is poor and has a negative impact on the business and users. But how can we identify performance issues quickly to prevent negative effects? We cannot alert on every slow transaction, since there are always some. In addition, most operations teams have to maintain a large number of applications and are not familiar with all of them, so manually setting thresholds can be inaccurate, quite painful and time-consuming.

The industry has come up with a solution called Automatic Baselining. Baselining calculates out the "normal" performance and only alerts us when an application slows down or produces more errors than usual. Most approaches rely on averages and standard deviations.

Without going into statistical details, this approach again assumes that the response times are distributed over a bell curve:

The Standard Deviation represents 33% of all transactions with the mean as the middle. 2xStandard Deviation represents 66% and thus the majority, everything outside could be considered an outlier. However most real world scenarios are not bell curved...

Typically, transactions that are outside two times standard deviation are treated as slow and captured for analysis. An alert is raised if the average moves significantly. In a bell curve this would account for the slowest 16.5 percent (and you can of course adjust that); however; if the response time distribution does not represent a bell curve, it becomes inaccurate. We either end up with a lot of false positives (transactions that are a lot slower than the average but when looking at the curve lie within the norm) or we miss a lot of problems (false negatives). In addition if the curve is not a bell curve, then the average can differ a lot from the median; applying a standard deviation to such an average can lead to quite a different result than you would expect. To work around this problem these algorithms have many tunable variables and a lot of "hacks" for specific use cases.

Why I Love Percentiles
A percentile tells me which part of the curve I am looking at and how many transactions are represented by that metric. To visualize this look at the following chart:

This chart shows the 50th and 90th percentile along with the average of the same transaction. It shows that the average is influenced far mor heavily by the 90th, thus by outliers and not by the bulk of the transactions

The green line represents the average. As you can see it is very volatile. The other two lines represent the 50th and 90th percentile. As we can see the 50th percentile (or median) is rather stable but has a couple of jumps. These jumps represent real performance degradation for the majority (50%) of the transactions. The 90th percentile (this is the start of the "tail") is a lot more volatile, which means that the outliers slowness depends on data or user behavior. What's important here is that the average is heavily influenced (dragged) by the 90th percentile, the tail, rather than the bulk of the transactions.

If the 50th percentile (median) of a response time is 500ms that means that 50% of my transactions are either as fast or faster than 500ms. If the 90th percentile of the same transaction is at 1000ms it means that 90% are as fast or faster and only 10% are slower. The average in this case could either be lower than 500ms (on a heavy front curve), a lot higher (long tail) or somewhere in between. A percentile gives me a much better sense of my real world performance, because it shows me a slice of my response time curve.

For exactly that reason percentiles are perfect for automatic baselining. If the 50th percentile moves from 500ms to 600ms I know that 50% of my transactions suffered a 20% performance degradation. You need to react to that.

In many cases we see that the 75th or 90th percentile does not change at all in such a scenario. This means the slow transactions didn't get any slower, only the normal ones did. Depending on how long your tail is the average might not have moved at all in such a scenario.!

In other cases we see the 98th percentile degrading from 1s to 1.5 seconds while the 95th is stable at 900ms. This means that your application as a whole is stable, but a few outliers got worse, nothing to worry about immediately. Percentile-based alerts do not suffer from false positives, are a lot less volatile and don't miss any important performance degradations! Consequently a baselining approach that uses percentiles does not require a lot of tuning variables to work effectively.

The screenshot below shows the Median (50th Percentile) for a particular transaction jumping from about 50ms to about 500ms and triggering an alert as it is significantly above the calculated baseline (green line). The chart labeled "Slow Response Time" on the other hand shows the 90th percentile for the same transaction. These "outliers" also show an increase in response time but not significant enough to trigger an alert.

Here we see an automatic baselining dashboard with a violation at the 50th percentile. The violation is quite clear, at the same time the 90th percentile (right upper chart) does not violate. Because the outliers are so much slower than the bulk of the transaction an average would have been influenced by them and would not have have reacted quite as dramatically as the 50th percentile. We might have missed this clear violation!

How Can We Use Percentiles for Tuning?
Percentiles are also great for tuning, and giving your optimizations a particular goal. Let's say that something within my application is too slow in general and I need to make it faster. In this case I want to focus on bringing down the 90th percentile. This would ensure sure that the overall response time of the application goes down. In other cases I have unacceptably long outliers I want to focus on bringing down response time for transactions beyond the 98th or 99th percentile (only outliers). We see a lot of applications that have perfectly acceptable performance for the 90th percentile, with the 98th percentile being magnitudes worse.

In throughput oriented applications on the other hand I would want to make the majority of my transactions very fast, while accepting that an optimization makes a few outliers slower. I might therefore make sure that the 75th percentile goes down while trying to keep the 90th percentile stable or not getting a lot worse.

I could not make the same kind of observations with averages, minimum and maximum, but with percentiles they are very easy indeed.

Averages are ineffective because they are too simplistic and one-dimensional. Percentiles are a really great and easy way of understanding the real performance characteristics of your application. They also provide a great basis for automatic baselining, behavioral learning and optimizing your application with a proper focus. In short, percentiles are great!

More Stories By Michael Kopp

Michael Kopp has over 12 years of experience as an architect and developer in the Enterprise Java space. Before coming to CompuwareAPM dynaTrace he was the Chief Architect at GoldenSource, a major player in the EDM space. In 2009 he joined dynaTrace as a technology strategist in the center of excellence. He specializes application performance management in large scale production environments with special focus on virtualized and cloud environments. His current focus is how to effectively leverage BigData Solutions and how these technologies impact and change the application landscape.

Comments (1) View Comments

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.

Most Recent Comments
rtalexander 11/21/12 12:58:00 AM EST

Hey, could you post a reference or two that covers the theory and/or practicalities of the approach you describe?


@MicroservicesExpo Stories
DevOps is being widely accepted (if not fully adopted) as essential in enterprise IT. But as Enterprise DevOps gains maturity, expands scope, and increases velocity, the need for data-driven decisions across teams becomes more acute. DevOps teams in any modern business must wrangle the ‘digital exhaust’ from the delivery toolchain, "pervasive" and "cognitive" computing, APIs and services, mobile devices and applications, the Internet of Things, and now even blockchain. In this power panel at @...
This is a no-hype, pragmatic post about why I think you should consider architecting your next project the way SOA and/or microservices suggest. No matter if it’s a greenfield approach or if you’re in dire need of refactoring. Please note: considering still keeps open the option of not taking that approach. After reading this, you will have a better idea about whether building multiple small components instead of a single, large component makes sense for your project. This post assumes that you...
What do dependency resolution, situational awareness, and superheroes have in common? Meet Chris Corriere, a DevOps/Software Engineer at Autotrader, speaking on creative ways to maximize usage of all of the above. Mark Miller, Community Advocate and senior storyteller at Sonatype, caught up with Chris to learn more about what his team is up to.
@DevOpsSummit has been named the ‘Top DevOps Influencer' by iTrend. iTrend processes millions of conversations, tweets, interactions, news articles, press releases, blog posts - and extract meaning form them and analyzes mobile and desktop software platforms used to communicate, various metadata (such as geo location), and automation tools. In overall placement, @DevOpsSummit ranked as the number one ‘DevOps Influencer' followed by @CloudExpo at third, and @MicroservicesE at 24th.
JetBlue Airways uses virtual environments to reduce software development costs, centralize performance testing, and create a climate for continuous integration and real-time monitoring of mobile applications. The next BriefingsDirect Voice of the Customer performance engineering case study discussion examines how JetBlue Airways in New York uses virtual environments to reduce software development costs, centralize performance testing, and create a climate for continuous integration and real-tim...
At its core DevOps is all about collaboration. The lines of communication must be opened and it takes some effort to ensure that they stay that way. It’s easy to pay lip service to trends and talk about implementing new methodologies, but without action, real benefits cannot be realized. Success requires planning, advocates empowered to effect change, and, of course, the right tooling. To bring about a cultural shift it’s important to share challenges. In simple terms, ensuring that everyone k...
SYS-CON Events announced today that SoftNet Solutions will exhibit at the 19th International Cloud Expo, which will take place on November 1–3, 2016, at the Santa Clara Convention Center in Santa Clara, CA. SoftNet Solutions specializes in Enterprise Solutions for Hadoop and Big Data. It offers customers the most open, robust, and value-conscious portfolio of solutions, services, and tools for the shortest route to success with Big Data. The unique differentiator is the ability to architect and ...
A completely new computing platform is on the horizon. They’re called Microservers by some, ARM Servers by others, and sometimes even ARM-based Servers. No matter what you call them, Microservers will have a huge impact on the data center and on server computing in general. Although few people are familiar with Microservers today, their impact will be felt very soon. This is a new category of computing platform that is available today and is predicted to have triple-digit growth rates for some ...
As the world moves toward more DevOps and Microservices, application deployment to the cloud ought to become a lot simpler. The Microservices architecture, which is the basis of many new age distributed systems such as OpenStack, NetFlix and so on, is at the heart of Cloud Foundry - a complete developer-oriented Platform as a Service (PaaS) that is IaaS agnostic and supports vCloud, OpenStack and AWS. Serverless computing is revolutionizing computing. In his session at 19th Cloud Expo, Raghav...
SYS-CON Events announced today that Transparent Cloud Computing (T-Cloud) Consortium will exhibit at the 19th International Cloud Expo®, which will take place on November 1–3, 2016, at the Santa Clara Convention Center in Santa Clara, CA. The Transparent Cloud Computing Consortium (T-Cloud Consortium) will conduct research activities into changes in the computing model as a result of collaboration between "device" and "cloud" and the creation of new value and markets through organic data proces...
So you think you are a DevOps warrior, huh? Put your money (not really, it’s free) where your metrics are and prove it by taking The Ultimate DevOps Geek Quiz Challenge, sponsored by DevOps Summit. Battle through the set of tough questions created by industry thought leaders to earn your bragging rights and win some cool prizes.
In case you haven’t heard, the new hotness in app architectures is serverless. Mainly restricted to cloud environments (Amazon Lambda, Google Cloud Functions, Microsoft Azure Functions) the general concept is that you don’t have to worry about anything but the small snippets of code (functions) you write to do something when something happens. That’s an event-driven model, by the way, that should be very familiar to anyone who has taken advantage of a programmable proxy to do app or API routing ...
More and more companies are looking to microservices as an architectural pattern for breaking apart applications into more manageable pieces so that agile teams can deliver new features quicker and more effectively. What this pattern has done more than anything to date is spark organizational transformations, setting the foundation for future application development. In practice, however, there are a number of considerations to make that go beyond simply “build, ship, and run,” which changes ho...
Analysis of 25,000 applications reveals 6.8% of packages/components used included known defects. Organizations standardizing on components between 2 - 3 years of age can decrease defect rates substantially. Open source and third-party packages/components live at the heart of high velocity software development organizations. Today, an average of 106 packages/components comprise 80 - 90% of a modern application, yet few organizations have visibility into what components are used where.
SYS-CON Events announced today that Enzu will exhibit at the 19th International Cloud Expo, which will take place on November 1–3, 2016, at the Santa Clara Convention Center in Santa Clara, CA. Enzu’s mission is to be the leading provider of enterprise cloud solutions worldwide. Enzu enables online businesses to use its IT infrastructure to their competitive advantage. By offering a suite of proven hosting and management services, Enzu wants companies to focus on the core of their online busine...
With emerging ideas, innovation, and talents, the lines between DevOps, release engineering, and even security are rapidly blurring. I invite you to sit down for a moment with Principle Consultant, J. Paul Reed, and listen to his take on what the intersection between these once individualized fields entails, and may even foreshadow.
In many organizations governance is still practiced by phase or stage gate peer review, and Agile projects are forced to accommodate, which leads to WaterScrumFall or worse. But governance criteria and policies are often very weak anyway, out of date or non-existent. Consequently governance is frequently a matter of opinion and experience, highly dependent upon the experience of individual reviewers. As we all know, a basic principle of Agile methods is delegation of responsibility, and ideally ...
Monitoring of Docker environments is challenging. Why? Because each container typically runs a single process, has its own environment, utilizes virtual networks, or has various methods of managing storage. Traditional monitoring solutions take metrics from each server and applications they run. These servers and applications running on them are typically very static, with very long uptimes. Docker deployments are different: a set of containers may run many applications, all sharing the resource...
When we talk about the impact of BYOD and BYOA and the Internet of Things, we often focus on the impact on data center architectures. That's because there will be an increasing need for authentication, for access control, for security, for application delivery as the number of potential endpoints (clients, devices, things) increases. That means scale in the data center. What we gloss over, what we skip, is that before any of these "things" ever makes a request to access an application it had to...
Virgil consists of an open-source encryption library, which implements Cryptographic Message Syntax (CMS) and Elliptic Curve Integrated Encryption Scheme (ECIES) (including RSA schema), a Key Management API, and a cloud-based Key Management Service (Virgil Keys). The Virgil Keys Service consists of a public key service and a private key escrow service.