Click here to close now.

Welcome!

XML Authors: Dana Gardner, Ruxit Blog, VictorOps Blog, Carmen Gonzalez, Automic Blog

Related Topics: XML, Java, SOA & WOA

XML: Article

Troubleshooting Response Time Problems

Why you cannot trust your system metrics

Production Monitoring is about ensuring the stability and health of our system, that also includes the application. A lot of times we encounter production systems that concentrate on System Monitoring, under the assumption that a stable system leads to stable and healthy applications. So let’s see what System Monitoring can tell us about our Application.

Let’s take a very simple two-tier Web Application:

A simple two tier web application

A simple two tier web application

This is a simple multi-tier eCommerce solution. Users are concerned about bad performance when they do a search. Let's see what we can find out about it if performance is not satisfactory. We start by looking at a couple of simple metrics.

CPU Utilization
The best known operating system metric is CPU utilization, but it is also the most misunderstood. This metric tells us how much time the CPU spent executing code in the last interval and how much more it could execute theoretically. Like all other utilization measures it tells us something about the capacity, but not about health, stability or even performance. Simply put: 99% CPU utilization can either be optimal or indicate impeding disaster depending on the application.

The CPU Usage of the two tiers

The CPU charts show no shortage on either tier

Let's look at our setup. We see that the CPU utilization is well below 100%, so we do have capacity left. But does that mean the machine or the application can be considered healthy? Let’s look at another measure that is better suited for the job, the Load Average  (System\Processor QueueLength on Windows). The Load Average tells us how many threads or processes are currently executed or waiting to get CPU time.

Unix Top Output: load average: 1.31, 1.13, 1.10

Linux systems display three sliding load averages for the last one, five and 15 minutes. The output above shows that in the last minute there were on average 1.3 processes that needed a CPU core at the same time.

If the Load Average is higher than the number of cores in the system we should either see near 100% CPU utilization, or the system has to wait for other resources and cannot max out the CPU. Examples would be Swapping or other I/O related tasks. So the Load Average tells us if we should trust the CPU usage on the one hand and if the machine is overloaded on the other. It does not tell us how well the application itself is performing, but whether the shortage of CPU might impact it negatively. If we do notice a problem we can identify the application that is causing the issue, but not why it is causing it.

In our case we see that neither the load average nor the CPU usage shines any light on our performance issue. If it were to show high CPU utilization or a high load average we could assume that the shortage in CPU is a problem, but we could not be certain.

Memory Usage
Used memory is monitored because the lack of memory will lead to system instability. An important fact to note is that Unix and Linux operating systems will most always show close to 100% memory utilization over time. They fill the memory up with buffers and caches which get discarded, as opposed to swapped out, if that memory is needed otherwise. In order to get the "real" memory usage we need subtract these. In Linux we can do by using the free command.

Memory Usage on the two systems

Memory Usage on the two systems, neither is suffering memory problems

If we do not have enough memory we can try to identify which application consumes the most by looking at the resident memory usage of a process. Once identified we will have to use other means to identify why the process uses up the memory and whether this is ok. When we look towards memory regarding Java/.NET performance we have to make sure that the application itself is never swapped out. This is especially important because  Java accesses all its memory in a random-access fashion and if a portion were to be swapped out it would have serve performance penalties. We can monitor this via swapping measures on the process itself. So what we can learn here is whether the shortage of memory has a negative impact on application performance. As this is not the case, we are tempted to ignore memory as the issue.

We could look at other measures like network or disk, but in all cases the same thing would be true, the shortage of a resource might have impact, but we cannot say for sure. And if we don't find a shortage it does not necessarily mean that everything is fine.

Database
An especially good example of this problem is the database. Very often the database is considered the source of all performance problems, at least by the application people. From a DBA's and operations point of view the database is often running fine though.  Their reasoning is simple enough, the database is not running out of any resources, there are no especially long running or CPU consuming statements or processes running and most statements execute quite fast. So the database can not be the problem.

Let's look at this from an application point of view

Looking at the Application
As users are reporting performance problems the first thing that we do is to look at the response time and its distribution within our system.

The overall distribution in our web application

The overall distribution in our system does not show any particular bottleneck

At first glance we don't see anything particularly interesting when looking at the whole system. As users are complaining about specific requests lets go ahead and look at these in particular:

Response time distribution of the search

The response time distribution of the specific request shows a bottleneck in the backend and a lot of database calls for each and every search request

We see that the majority of the response time lies in the backend and the database layer. That the database contributes a major portion to the response time does not mean however that the DBA was wrong. We see that every single search executes 416 statements on average! That means that every statement is executing in under one millisecond and this is fast enough from the database point of view. The problem really lies within the application and its usage of the database. Let's look at the backend next.

Heap usage and GC activity on the backend

The Heap Usage and GC activity chart shows a lot of GC runs, but does it have negative impact?

Looking at the JVM we immediately see that it does execute a lot of garbage collection (the red spikes), as you would probably see in every monitoring tool. Although this gives us a strong suspicion, we do not know how this is affecting our users. So let's look at that impact:

GC Runtime suspensions that have an impact on the search

These are the runtime suspensions that directly impact the search. It is considerable but still amounts to only 10% of the response time

A single transaction is hit by garbage collection several times and if we do the math we find out that garbage collection contributes 10% to the response time. While that is considerable it would not have made sense to spend a lot of time on tuning it just now. Even if we get it down to half it would only have saved us 5% of the response time. So while monitoring garbage collection is important, we should always analyze the impact before we jump to conclusions.

So let's take a deeper look at where that particular transaction is spending time on the backend. To do this we need to have application centric monitoring in place which we can then use to isolate the root cause.

Response time distribution of the search within the backend

The detailed response time distribution of the search within the backend shows two main problems: too many EJB calls and a very slow doPost method

With the right measure points within our application we immediately see the root causes of the response time problem. At first we see that the WebService call done by the search takes up a large portion of the response time. It is also the largest CPU hotspot within that call. So while the host is not suffering CPU problems, we are in fact consuming a lot of it in that particular transaction. Secondly we see that an awful lot of EJB calls are done which in turn leads to the many database calls that we have already noticed.

That means we have identified a small memory-related issue; although there are no memory problems noticeable if we were to look only at system monitoring. We also found that we have a CPU hotspot, but the machine itself does not have a CPU problem. And finally we found that the biggest issue is squarely within the application; too many database and EJB calls, which we cannot see on a system monitoring level at all.

Conclusion
System metrics do a very good job at describing the environment, after all that is what they are meant for. If the environment itself has resource shortages we can almost assume that this has a negative impact on the applications, but we cannot be sure. If there is no obvious shortage this does not, however, imply that the application is running smoothly. A healthy and stable environment does not guarantee a healthy, stable and performing application.

Similar to the system, the application needs to be monitored in detail and with application-specific metrics in order to ensure its health and stability. There is no universal rule as to what these metrics are, but they should enable us to describe the health, stability and performance of the application itself.

Related reading:

  1. The impact of Garbage Collection on Java performance // In my last post I explained what a major...
  2. Top 10 Performance Problems taken from Zappos, Monster, Thomson and Co For a recent edition of the Swiss Computerworld Magazine we...
  3. Top 10 Client-Side Performance Problems in Web 2.0 Inspired by the Top 10 Performance Problems post which focuses...
  4. Real Life Ajax Troubleshooting Guide One of our clients occasionally runs into the following problem...
  5. Presenting Top Web 2.0 Performance Problems at WebTechCon 2010 in Mainz, Germany WebTech Conference 2010 takes place in Mainz, Germany from October...

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.

@ThingsExpo Stories
The true value of the Internet of Things (IoT) lies not just in the data, but through the services that protect the data, perform the analysis and present findings in a usable way. With many IoT elements rooted in traditional IT components, Big Data and IoT isn’t just a play for enterprise. In fact, the IoT presents SMBs with the prospect of launching entirely new activities and exploring innovative areas. CompTIA research identifies several areas where IoT is expected to have the greatest impact.
Wearable devices have come of age. The primary applications of wearables so far have been "the Quantified Self" or the tracking of one's fitness and health status. We propose the evolution of wearables into social and emotional communication devices. Our BE(tm) sensor uses light to visualize the skin conductance response. Our sensors are very inexpensive and can be massively distributed to audiences or groups of any size, in order to gauge reactions to performances, video, or any kind of presentation. In her session at @ThingsExpo, Jocelyn Scheirer, CEO & Founder of Bionolux, will discuss ho...
SYS-CON Events announced today that GENBAND, a leading developer of real time communications software solutions, has been named “Silver Sponsor” of SYS-CON's WebRTC Summit, which will take place on June 9-11, 2015, at the Javits Center in New York City, NY. The GENBAND team will be on hand to demonstrate their newest product, Kandy. Kandy is a communications Platform-as-a-Service (PaaS) that enables companies to seamlessly integrate more human communications into their Web and mobile applications - creating more engaging experiences for their customers and boosting collaboration and productiv...
Roberto Medrano, Executive Vice President at SOA Software, had reached 30,000 page views on his home page - http://RobertoMedrano.SYS-CON.com/ - on the SYS-CON family of online magazines, which includes Cloud Computing Journal, Internet of Things Journal, Big Data Journal, and SOA World Magazine. He is a recognized executive in the information technology fields of SOA, internet security, governance, and compliance. He has extensive experience with both start-ups and large companies, having been involved at the beginning of four IT industries: EDA, Open Systems, Computer Security and now SOA.
From telemedicine to smart cars, digital homes and industrial monitoring, the explosive growth of IoT has created exciting new business opportunities for real time calls and messaging. In his session at @ThingsExpo, Ivelin Ivanov, CEO and Co-Founder of Telestax, shared some of the new revenue sources that IoT created for Restcomm – the open source telephony platform from Telestax. Ivelin Ivanov is a technology entrepreneur who founded Mobicents, an Open Source VoIP Platform, to help create, deploy, and manage applications integrating voice, video and data. He is the co-founder of TeleStax, a...
The industrial software market has treated data with the mentality of “collect everything now, worry about how to use it later.” We now find ourselves buried in data, with the pervasive connectivity of the (Industrial) Internet of Things only piling on more numbers. There’s too much data and not enough information. In his session at @ThingsExpo, Bob Gates, Global Marketing Director, GE’s Intelligent Platforms business, to discuss how realizing the power of IoT, software developers are now focused on understanding how industrial data can create intelligence for industrial operations. Imagine ...
Operational Hadoop and the Lambda Architecture for Streaming Data Apache Hadoop is emerging as a distributed platform for handling large and fast incoming streams of data. Predictive maintenance, supply chain optimization, and Internet-of-Things analysis are examples where Hadoop provides the scalable storage, processing, and analytics platform to gain meaningful insights from granular data that is typically only valuable from a large-scale, aggregate view. One architecture useful for capturing and analyzing streaming data is the Lambda Architecture, representing a model of how to analyze rea...
SYS-CON Events announced today that Vitria Technology, Inc. will exhibit at SYS-CON’s @ThingsExpo, which will take place on June 9-11, 2015, at the Javits Center in New York City, NY. Vitria will showcase the company’s new IoT Analytics Platform through live demonstrations at booth #330. Vitria’s IoT Analytics Platform, fully integrated and powered by an operational intelligence engine, enables customers to rapidly build and operationalize advanced analytics to deliver timely business outcomes for use cases across the industrial, enterprise, and consumer segments.
When it comes to the Internet of Things, hooking up will get you only so far. If you want customers to commit, you need to go beyond simply connecting products. You need to use the devices themselves to transform how you engage with every customer and how you manage the entire product lifecycle. In his session at @ThingsExpo, Sean Lorenz, Technical Product Manager for Xively at LogMeIn, will show how “product relationship management” can help you leverage your connected devices and the data they generate about customer usage and product performance to deliver extremely compelling and reliabl...
The explosion of connected devices / sensors is creating an ever-expanding set of new and valuable data. In parallel the emerging capability of Big Data technologies to store, access, analyze, and react to this data is producing changes in business models under the umbrella of the Internet of Things (IoT). In particular within the Insurance industry, IoT appears positioned to enable deep changes by altering relationships between insurers, distributors, and the insured. In his session at @ThingsExpo, Michael Sick, a Senior Manager and Big Data Architect within Ernst and Young's Financial Servi...
SYS-CON Events announced today that Open Data Centers (ODC), a carrier-neutral colocation provider, will exhibit at SYS-CON's 16th International Cloud Expo®, which will take place June 9-11, 2015, at the Javits Center in New York City, NY. Open Data Centers is a carrier-neutral data center operator in New Jersey and New York City offering alternative connectivity options for carriers, service providers and enterprise customers.
SYS-CON Events announced today that CodeFutures, a leading supplier of database performance tools, has been named a “Sponsor” of SYS-CON's 16th International Cloud Expo®, which will take place on June 9–11, 2015, at the Javits Center in New York, NY. CodeFutures is an independent software vendor focused on providing tools that deliver database performance tools that increase productivity during database development and increase database performance and scalability during production.
The IoT market is projected to be $1.9 trillion tidal wave that’s bigger than the combined market for smartphones, tablets and PCs. While IoT is widely discussed, what not being talked about are the monetization opportunities that are created from ubiquitous connectivity and the ensuing avalanche of data. While we cannot foresee every service that the IoT will enable, we should future-proof operations by preparing to monetize them with extremely agile systems.
There’s Big Data, then there’s really Big Data from the Internet of Things. IoT is evolving to include many data possibilities like new types of event, log and network data. The volumes are enormous, generating tens of billions of logs per day, which raise data challenges. Early IoT deployments are relying heavily on both the cloud and managed service providers to navigate these challenges. Learn about IoT, Big Data and deployments processing massive data volumes from wearables, utilities and other machines.
The explosion of connected devices / sensors is creating an ever-expanding set of new and valuable data. In parallel the emerging capability of Big Data technologies to store, access, analyze, and react to this data is producing changes in business models under the umbrella of the Internet of Things (IoT). In particular within the Insurance industry, IoT appears positioned to enable deep changes by altering relationships between insurers, distributors, and the insured. In his session at @ThingsExpo, Michael Sick, a Senior Manager and Big Data Architect within Ernst and Young's Financial Servi...
The major cloud platforms defy a simple, side-by-side analysis. Each of the major IaaS public-cloud platforms offers their own unique strengths and functionality. Options for on-site private cloud are diverse as well, and must be designed and deployed while taking existing legacy architecture and infrastructure into account. Then the reality is that most enterprises are embarking on a hybrid cloud strategy and programs. In this Power Panel at 15th Cloud Expo (http://www.CloudComputingExpo.com), moderated by Ashar Baig, Research Director, Cloud, at Gigaom Research, Nate Gordon, Director of T...
“In the past year we've seen a lot of stabilization of WebRTC. You can now use it in production with a far greater degree of certainty. A lot of the real developments in the past year have been in things like the data channel, which will enable a whole new type of application," explained Peter Dunkley, Technical Director at Acision, in this SYS-CON.tv interview at @ThingsExpo, held Nov 4–6, 2014, at the Santa Clara Convention Center in Santa Clara, CA.
SYS-CON Events announced today that Intelligent Systems Services will exhibit at SYS-CON's 16th International Cloud Expo®, which will take place on June 9-11, 2015, at the Javits Center in New York City, NY. Established in 1994, Intelligent Systems Services Inc. is located near Washington, DC, with representatives and partners nationwide. ISS’s well-established track record is based on the continuous pursuit of excellence in designing, implementing and supporting nationwide clients’ mission-critical systems. ISS has completed many successful projects in Healthcare, Commercial, Manufacturing, ...
PubNub on Monday has announced that it is partnering with IBM to bring its sophisticated real-time data streaming and messaging capabilities to Bluemix, IBM’s cloud development platform. “Today’s app and connected devices require an always-on connection, but building a secure, scalable solution from the ground up is time consuming, resource intensive, and error-prone,” said Todd Greene, CEO of PubNub. “PubNub enables web, mobile and IoT developers building apps on IBM Bluemix to quickly add scalable realtime functionality with minimal effort and cost.”
Sensor-enabled things are becoming more commonplace, precursors to a larger and more complex framework that most consider the ultimate promise of the IoT: things connecting, interacting, sharing, storing, and over time perhaps learning and predicting based on habits, behaviors, location, preferences, purchases and more. In his session at @ThingsExpo, Tom Wesselman, Director of Communications Ecosystem Architecture at Plantronics, will examine the still nascent IoT as it is coalescing, including what it is today, what it might ultimately be, the role of wearable tech, and technology gaps stil...