Microservices Expo Authors: Elizabeth White, Mehdi Daoudi, Pat Romanski, Flint Brenton, Gordon Haff

Related Topics: Java IoT, Microservices Expo, Adobe Flex, Machine Learning , 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We recognize that the average no longer represents the bulk of the transactions but can be a lot higher than the median.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
The dynamic nature of the cloud means that change is a constant when it comes to modern cloud-based infrastructure. Delivering modern applications to end users, therefore, is a constantly shifting challenge. Delivery automation helps IT Ops teams ensure that apps are providing an optimal end user experience over hybrid-cloud and multi-cloud environments, no matter what the current state of the infrastructure is. To employ a delivery automation strategy that reflects your business rules, making r...
"We started a Master of Science in business analytics - that's the hot topic. We serve the business community around San Francisco so we educate the working professionals and this is where they all want to be," explained Judy Lee, Associate Professor and Department Chair at Golden Gate University, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
For over a decade, Application Programming Interface or APIs have been used to exchange data between multiple platforms. From social media to news and media sites, most websites depend on APIs to provide a dynamic and real-time digital experience. APIs have made its way into almost every device and service available today and it continues to spur innovations in every field of technology. There are multiple programming languages used to build and run applications in the online world. And just li...
There is a huge demand for responsive, real-time mobile and web experiences, but current architectural patterns do not easily accommodate applications that respond to events in real time. Common solutions using message queues or HTTP long-polling quickly lead to resiliency, scalability and development velocity challenges. In his session at 21st Cloud Expo, Ryland Degnan, a Senior Software Engineer on the Netflix Edge Platform team, will discuss how by leveraging a reactive stream-based protocol,...
The general concepts of DevOps have played a central role advancing the modern software delivery industry. With the library of DevOps best practices, tips and guides expanding quickly, it can be difficult to track down the best and most accurate resources and information. In order to help the software development community, and to further our own learning, we reached out to leading industry analysts and asked them about an increasingly popular tenet of a DevOps transformation: collaboration.
We call it DevOps but much of the time there’s a lot more discussion about the needs and concerns of developers than there is about other groups. There’s a focus on improved and less isolated developer workflows. There are many discussions around collaboration, continuous integration and delivery, issue tracking, source code control, code review, IDEs, and xPaaS – and all the tools that enable those things. Changes in developer practices may come up – such as developers taking ownership of code ...
Cloud Governance means many things to many people. Heck, just the word cloud means different things depending on who you are talking to. While definitions can vary, controlling access to cloud resources is invariably a central piece of any governance program. Enterprise cloud computing has transformed IT. Cloud computing decreases time-to-market, improves agility by allowing businesses to adapt quickly to changing market demands, and, ultimately, drives down costs.
Modern software design has fundamentally changed how we manage applications, causing many to turn to containers as the new virtual machine for resource management. As container adoption grows beyond stateless applications to stateful workloads, the need for persistent storage is foundational - something customers routinely cite as a top pain point. In his session at @DevOpsSummit at 21st Cloud Expo, Bill Borsari, Head of Systems Engineering at Datera, explored how organizations can reap the bene...
How is DevOps going within your organization? If you need some help measuring just how well it is going, we have prepared a list of some key DevOps metrics to track. These metrics can help you understand how your team is doing over time. The word DevOps means different things to different people. Some say it a culture and every vendor in the industry claims that their tools help with DevOps. Depending on how you define DevOps, some of these metrics may matter more or less to you and your team.
"CA has been doing a lot of things in the area of DevOps. Now we have a complete set of tool sets in order to enable customers to go all the way from planning to development to testing down to release into the operations," explained Aruna Ravichandran, Vice President of Global Marketing and Strategy at CA Technologies, in this SYS-CON.tv interview at DevOps Summit at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
"We are an integrator of carrier ethernet and bandwidth to get people to connect to the cloud, to the SaaS providers, and the IaaS providers all on ethernet," explained Paul Mako, CEO & CTO of Massive Networks, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
"Grape Up leverages Cloud Native technologies and helps companies build software using microservices, and work the DevOps agile way. We've been doing digital innovation for the last 12 years," explained Daniel Heckman, of Grape Up in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
"NetApp's vision is how we help organizations manage data - delivering the right data in the right place, in the right time, to the people who need it, and doing it agnostic to what the platform is," explained Josh Atwell, Developer Advocate for NetApp, in this SYS-CON.tv interview at 20th Cloud Expo, held June 6-8, 2017, at the Javits Center in New York City, NY.
"Outscale was founded in 2010, is based in France, is a strategic partner to Dassault Systémes and has done quite a bit of work with divisions of Dassault," explained Jackie Funk, Digital Marketing exec at Outscale, in this SYS-CON.tv interview at 20th Cloud Expo, held June 6-8, 2017, at the Javits Center in New York City, NY.
"I focus on what we are calling CAST Highlight, which is our SaaS application portfolio analysis tool. It is an extremely lightweight tool that can integrate with pretty much any build process right now," explained Andrew Siegmund, Application Migration Specialist for CAST, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
Let's do a visualization exercise. Imagine it's December 31, 2018, and you're ringing in the New Year with your friends and family. You think back on everything that you accomplished in the last year: your company's revenue is through the roof thanks to the success of your product, and you were promoted to Lead Developer. 2019 is poised to be an even bigger year for your company because you have the tools and insight to scale as quickly as demand requires. You're a happy human, and it's not just...
The enterprise data storage marketplace is poised to become a battlefield. No longer the quiet backwater of cloud computing services, the focus of this global transition is now going from compute to storage. An overview of recent storage market history is needed to understand why this transition is important. Before 2007 and the birth of the cloud computing market we are witnessing today, the on-premise model hosted in large local data centers dominated enterprise storage. Key marketplace play...
Cavirin Systems has just announced C2, a SaaS offering designed to bring continuous security assessment and remediation to hybrid environments, containers, and data centers. Cavirin C2 is deployed within Amazon Web Services (AWS) and features a flexible licensing model for easy scalability and clear pay-as-you-go pricing. Although native to AWS, it also supports assessment and remediation of virtual or container instances within Microsoft Azure, Google Cloud Platform (GCP), or on-premise. By dr...
With continuous delivery (CD) almost always in the spotlight, continuous integration (CI) is often left out in the cold. Indeed, it's been in use for so long and so widely, we often take the model for granted. So what is CI and how can you make the most of it? This blog is intended to answer those questions. Before we step into examining CI, we need to look back. Software developers often work in small teams and modularity, and need to integrate their changes with the rest of the project code b...
Kubernetes is an open source system for automating deployment, scaling, and management of containerized applications. Kubernetes was originally built by Google, leveraging years of experience with managing container workloads, and is now a Cloud Native Compute Foundation (CNCF) project. Kubernetes has been widely adopted by the community, supported on all major public and private cloud providers, and is gaining rapid adoption in enterprises. However, Kubernetes may seem intimidating and complex ...