|By Michael Kopp||
|December 12, 2012 07:30 AM EST||
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We recognize that the average no longer represents the bulk of the transactions but can be a lot higher than the median.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!
|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?
[slides] Storage for Docker Containers By @OnModulus | @DevOpsSummit #DevOps #Docker #Containers #Microservices
Learn how to solve the problem of keeping files in sync between multiple Docker containers. In his session at 16th Cloud Expo, Aaron Brongersma, Senior Infrastructure Engineer at Modulus, discussed using rsync, GlusterFS, EBS and Bit Torrent Sync. He broke down the tools that are needed to help create a seamless user experience. In the end, can we have an environment where we can easily move Docker containers, servers, and volumes without impacting our applications? He shared his results so yo...
Jul. 31, 2015 11:45 PM EDT Reads: 783
Modern DevOps Tool Kit By @Logentries and @NewRelic | @DevOpsSummit #DevOps #Containers #Microservices
Auto-scaling environments, micro-service architectures and globally-distributed teams are just three common examples of why organizations today need automation and interoperability more than ever. But is interoperability something we simply start doing, or does it require a reexamination of our processes? And can we really improve our processes without first making interoperability a requirement for how we choose our tools?
Jul. 31, 2015 11:15 PM EDT Reads: 425
Cloud Migration Management (CMM) refers to the best practices for planning and managing migration of IT systems from a legacy platform to a Cloud Provider through a combination professional services consulting and software tools. A Cloud migration project can be a relatively simple exercise, where applications are migrated ‘as is’, to gain benefits such as elastic capacity and utility pricing, but without making any changes to the application architecture, software development methods or busine...
Jul. 31, 2015 10:00 PM EDT Reads: 1,345
The Software Defined Data Center (SDDC), which enables organizations to seamlessly run in a hybrid cloud model (public + private cloud), is here to stay. IDC estimates that the software-defined networking market will be valued at $3.7 billion by 2016. Security is a key component and benefit of the SDDC, and offers an opportunity to build security 'from the ground up' and weave it into the environment from day one. In his session at 16th Cloud Expo, Reuven Harrison, CTO and Co-Founder of Tufin,...
Jul. 31, 2015 03:00 PM EDT Reads: 495
Jul. 31, 2015 02:00 PM EDT Reads: 290
You often hear the two titles of "DevOps" and "Immutable Infrastructure" used independently. In his session at DevOps Summit, John Willis, Technical Evangelist for Docker, covered the union between the two topics and why this is important. He provided an overview of Immutable Infrastructure then showed how an Immutable Continuous Delivery pipeline can be applied as a best practice for "DevOps." He ended the session with some interesting case study examples.
Jul. 31, 2015 02:00 PM EDT Reads: 194
Approved this February by the Internet Engineering Task Force (IETF), HTTP/2 is the first major update to HTTP since 1999, when HTTP/1.1 was standardized. Designed with performance in mind, one of the biggest goals of HTTP/2 implementation is to decrease latency while maintaining a high-level compatibility with HTTP/1.1. Though not all testing activities will be impacted by the new protocol, it's important for testers to be aware of any changes moving forward.
Jul. 31, 2015 12:30 PM EDT Reads: 166
One of the ways to increase scalability of services – and applications – is to go “stateless.” The reasons for this are many, but in general by eliminating the mapping between a single client and a single app or service instance you eliminate the need for resources to manage state in the app (overhead) and improve the distributability (I can make up words if I want) of requests across a pool of instances. The latter occurs because sessions don’t need to hang out and consume resources that could ...
Jul. 31, 2015 11:45 AM EDT Reads: 198
Alibaba, the world’s largest ecommerce provider, has pumped over a $1 billion into its subsidiary, Aliya, a cloud services provider. This is perhaps one of the biggest moments in the global Cloud Wars that signals the entry of China into the main arena. Here is why this matters. The cloud industry worldwide is being propelled into fast growth by tremendous demand for cloud computing services. Cloud, which is highly scalable and offers low investment and high computational capabilities to end us...
Jul. 31, 2015 11:00 AM EDT Reads: 120
The Internet of Things. Cloud. Big Data. Real-Time Analytics. To those who do not quite understand what these phrases mean (and let’s be honest, that’s likely to be a large portion of the world), words like “IoT” and “Big Data” are just buzzwords. The truth is, the Internet of Things encompasses much more than jargon and predictions of connected devices. According to Parker Trewin, Senior Director of Content and Communications of Aria Systems, “IoT is big news because it ups the ante: Reach out ...
Jul. 31, 2015 07:00 AM EDT Reads: 405
Where the Network Got Invited to the Party By @LMacVittie | @DevOpsSummit #DevOps #Docker #Containers #Microservices
At DevOps Summit NY there’s been a whole lot of talk about not just DevOps, but containers, IoT, and microservices. Sessions focused not just on the cultural shift needed to grow at scale with a DevOps approach, but also made sure to include the network ”plumbing” needed to ensure success as applications decompose into the microservice architectures enabling rapid growth and support for the Internet of (Every)Things.
Jul. 30, 2015 08:15 PM EDT Reads: 1,774
Designing the IT Architecture of the Future with Adrian Cockcroft | @DevOpsSummit #DevOps #Docker #Containers #Microservices
Our guest on the podcast this week is Adrian Cockcroft, Technology Fellow at Battery Ventures. We discuss what makes Docker and Netflix highly successful, especially through their use of well-designed IT architecture and DevOps.
Jul. 30, 2015 08:00 PM EDT Reads: 786
[slides] A New Architecture for the Internet of Things By @JKirklan | @ThingsExpo @RedHatNews #IoT #M2M #InternetOfThings
Explosive growth in connected devices. Enormous amounts of data for collection and analysis. Critical use of data for split-second decision making and actionable information. All three are factors in making the Internet of Things a reality. Yet, any one factor would have an IT organization pondering its infrastructure strategy. How should your organization enhance its IT framework to enable an Internet of Things implementation? In his session at @ThingsExpo, James Kirkland, Red Hat's Chief Arch...
Jul. 30, 2015 07:30 PM EDT Reads: 1,403
This week, I joined SOASTA as Senior Vice President of Performance Analytics. Given my background in cloud computing and distributed systems operations — you may have read my blogs on CNET or GigaOm — this may surprise you, but I want to explain why this is the perfect time to take on this opportunity with this team. In fact, that’s probably the best way to break this down. To explain why I’d leave the world of infrastructure and code for the world of data and analytics, let’s explore the timing...
Jul. 30, 2015 05:45 PM EDT Reads: 384
Take the Long View with Digital Transformation By @IoT2040 | @ThingsExpo #IoT #M2M #API #Microservices #InternetOfThings
Digital Transformation is the ultimate goal of cloud computing and related initiatives. The phrase is certainly not a precise one, and as subject to hand-waving and distortion as any high-falutin' terminology in the world of information technology. Yet it is an excellent choice of words to describe what enterprise IT—and by extension, organizations in general—should be working to achieve. Digital Transformation means: handling all the data types being found and created in the organizat...
Jul. 30, 2015 05:00 PM EDT Reads: 1,095
[slides] Workloads and Public Cloud at @CloudExpo By @utollwi | @ProfitBricksUSA #DevOps #Containers #Microservices
Public Cloud IaaS started its life in the developer and startup communities and has grown rapidly to a $20B+ industry, but it still pales in comparison to how much is spent worldwide on IT: $3.6 trillion. In fact, there are 8.6 million data centers worldwide, the reality is many small and medium sized business have server closets and colocation footprints filled with servers and storage gear. While on-premise environment virtualization may have peaked at 75%, the Public Cloud has lagged in adop...
Jul. 30, 2015 04:00 PM EDT Reads: 2,211
SYS-CON Events announced today that HPM Networks will exhibit at the 17th International Cloud Expo®, which will take place on November 3–5, 2015, at the Santa Clara Convention Center in Santa Clara, CA. For 20 years, HPM Networks has been integrating technology solutions that solve complex business challenges. HPM Networks has designed solutions for both SMB and enterprise customers throughout the San Francisco Bay Area.
Jul. 30, 2015 03:45 PM EDT Reads: 459
MuleSoft has announced the findings of its 2015 Connectivity Benchmark Report on the adoption and business impact of APIs. The findings suggest traditional businesses are quickly evolving into "composable enterprises" built out of hundreds of connected software services, applications and devices. Most are embracing the Internet of Things (IoT) and microservices technologies like Docker. A majority are integrating wearables, like smart watches, and more than half plan to generate revenue with ...
Jul. 30, 2015 02:30 PM EDT Reads: 110
[session] DevOps State of Mind By @RedHatNews | @DevOpsSummit #DevOps #PaaS #Jenkins #Kubernetes #Docker
Rapid innovation, changing business landscapes, and new IT demands force businesses to make changes quickly. The DevOps approach is a way to increase business agility through collaboration, communication, and integration across different teams in the IT organization. In his session at DevOps Summit, Chris Van Tuin, Chief Technologist for the Western US at Red Hat, will discuss: The acceleration of application delivery for the business with DevOps
Jul. 30, 2015 12:45 PM EDT Reads: 1,122
Growth hacking is common for startups to make unheard-of progress in building their business. Career Hacks can help Geek Girls and those who support them (yes, that's you too, Dad!) to excel in this typically male-dominated world. Get ready to learn the facts: Is there a bias against women in the tech / developer communities? Why are women 50% of the workforce, but hold only 24% of the STEM or IT positions? Some beginnings of what to do about it! In her Opening Keynote at 16th Cloud Expo, S...
Jul. 30, 2015 12:00 PM EDT Reads: 2,061