Welcome!

Machine Learning Authors: Ed Featherston, Destiny Bertucci, Liz McMillan, Elizabeth White, Dan Blacharski

Related Topics: @DXWorldExpo, Java IoT, Machine Learning , @CloudExpo, Apache, SDN Journal

@DXWorldExpo: Article

So What? – Monitoring Hadoop Beyond Ganglia

Don’t just run Hadoop jobs at scale, run them efficiently and at scale

Over the last couple of months I have been talking to more and more customers who are either bringing their Hadoop clusters into production or have already done so and are now getting serious about operations. This leads to some interesting discussions about how to monitor Hadoop properly and one thing pops up quite often: Do they need anything beyond Ganglia? If yes, what should they do beyond it?

The Basics
As in every other system, monitoring in a Hadoop environment starts with the basics: System Metrics - CPU, Disk, Memory you know the drill. Of special importance in a Hadoop system is a well-balanced cluster; you don't want to have some nodes being much more (or less) utilized then others. Besides CPU and memory utilization, Disk utilization and of course I/O throughput is of high importance. After all the most likely bottleneck in a Big Data system is I/O - either with ingress (network and disk), moving data around (e.g., MapReduce shuffle on the network) and straightforward read/write to disk.

The problem in a Hadoop system is of course its size. Nothing new for us, some of our customers monitor well beyond 1000+ JVMs with CompuwareAPM. The "advantage" in a Hadoop system is its relative conformity - every node looks pretty much like the other. This is what Ganglia leverages.

Cluster Monitoring with Ganglia
What Ganglia is very good at is providing an overview over how a cluster is utilized. The load chart is particularly interesting:

This chart shows the CPU load on a 1000 Server cluster that has roughly 15.000 CPUs

It tells us the number of available cores in the system and the number of running processes (in theory a core can never handle more than one process at a time) and the 1-min load average. If the system is getting fully utilized the 1-min load average would approach the total number of CPUs. Another view on this is the well-known CPU utilization chart:

CPU Utilization over the last day. While the utilization stays well below 10% we see a lot of I/O wait spikes.

While the load chart gives a good overall impression of usage, the utilization tells us the story of how the CPUs are used. While typical CPU charts show a single server, Ganglia specializes in showing whole clusters (the picture shows CPU usage of a 1000 machine cluster). In the case of the depicted chart we see that the CPUs are experiencing a lot of I/O wait spikes, which points toward heavy disk I/O. Basically it seems the disk I/O is the reason that we cannot utilize our CPU better at these times. But in general our cluster is well underutilized in terms of CPU.

Trends are also easy to understand, as can be seen in this memory chart over a year.

Memory capacity and usage over a year

All this looks pretty good, so what is missing? The "so what" and "why" is what is missing. If my memory demand is growing, I have no way of knowing why it is growing. If the CPU chart tells me that I spend a lot of time waiting, it does not tell what to do, or why that is so? These questions are beyond the scope of Ganglia.

What about Hadoop specifics?
Ganglia also has a Hadoop plugin, which basically gives you access to all the usual Hadoop metrics (unfortunately a comprehensive list of Hadoop metrics is really hard to find, appreciate if somebody commented the link). There is a good explanation on what is interesting on Edward Caproli's page: JoinTheGrid. Basically you can use those metrics to monitor the capacity and usage trends of HDFS and the NameNodes and also how many jobs, mappers and reducers are running.

Capacity of the DataNodes over time

Capacity of the Name Nodes over time

The DataNode Operations give me an impression of I/O pressure on the Hadoop cluster

All these charts can of course be easily built in any modern monitoring or APM solution like CompuwareAPM, but Ganglia gives you a simple starting point; and it's Free as in Beer.

What's missing again is the so what? If my jobs are running a lot longer than yesterday, what should I do? Why do they run longer? A Hadoop expert might dig into 10 different charts around I/O and Network, spilling, look at log files among other things and try an educated guess as to what might be the problem. But we aren't all experts, neither do we have the time to dig into all of these metrics and log files all the time.

This is the reason that we and our customers are moving beyond Ganglia - to solve the "Why" and "So What" within time constraints.

Beyond the Basics #1 - Understanding Cluster Utilization
A use case that we get from customers is that they want to know which users or which pools (in case of the fair scheduler) are responsible for how much of the cluster utilization. LinkedIn just released White Elephant, a tool that parses MapReduce logs and builds some nice dashboards and shows you which of your users occupy how much of your cluster. This is of course based on log file analysis and thus okay for analysis but not for monitoring. With proper tools in place we can do the same thing in near real time.

The CPU Usage in the Hadoop Cluster on per User basis

In this example I wanted to monitor which user consumed how much of my Amazon EMR cluster. If we see a user or pool that occupies a lot of the cluster we can course also see which jobs are running and how much of the cluster they occupy.

The CPU Usage in the Hadoop Cluster on per Job basis

And this will also tell us if that job has always been there, and just uses a lot more resources now. This would be our cue to start analyzing what has changed.

Beyond the Basics #2 - Understanding why my jobs are slow(er)
If we want to understand why a job is slow we need to look at a high-level break down first.

In which phase of the map reduce do we spend the most time, or did we spend more time than yesterday? Understanding these timings in context with the respective job counters, like Map Input or Spilled Records, helps us understand why the phase took longer.

Overview of the time spent in different phases and the respective input/output counters

At this point we will already have a pretty good idea as to what happened. We either simply have more data to crunch (more input data) or a portion of the MapReduce job consumes more CPU (code change?) or we spill more records to disk (code change or Hadoop config change?). We might also detect an unbalanced cluster in the performance breakdown.

This job is executing nearly exclusively on a single node instead of distributing

In this case we want to check whether all the involved nodes processed the same amount of data

Here we see that there is a wide range from minimum, average to maximum on mapped input and output records. The data is not balanced

or if the difference can again be found in the code (different kinds of computations). If we are running against HBase we might of course have an issue with HBase performance or distribution.

At the beginning of the job only a single HBase region Server consumes CPU while all others remain idle

On the other hand, if a lot of mapping time is spent in the garbage collector then you should maybe invest in larger JVMs.

The Performance Breakdown of this particular job shows considerable time in GC suspension

If spilling data to disk is where we spend our time, we should take a closer look at that phase. It might turn out that all of our time is spent on disk wait.

If the Disk were the bottleneck we would see it on disk I/O here

Now if disk write is our bottleneck, then really the only thing that we can do is reduce the map output records. Adding a combiner will not reduce the disk write (it will actually increase it, read here). In other words combining only optimizes the shuffle phase, thus the amount of data sent over the network, but not spill time!!

And at the very detailed level we can look at single task executions and understand in detail what is really going on.

The detailed data about each Map, Reduce Task Atttempt as well as the spills and shuffles

Conclusion
Ganglia is a great tool for high-level monitoring of your Hadoop cluster utilization, but it is not enough. The fact that everybody is working on additional means to understand the Hadoop cluster (Hortonworks with Ambari, Cloudera with their Manager, LinkedIn with White Elephant, the Star Fish project...) shows that there is a lot more needed beyond simple monitoring. Even those more advanced monitoring tools are not always answering the "why" though, which is what we really need to do. This is where the Performance Management discipline can add a lot of value and really help you get the best out of your Hadoop cluster. In other words don't just run Hadoop jobs at scale, run them efficiently and at scale!

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 (0)

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.


@CloudExpo Stories
A strange thing is happening along the way to the Internet of Things, namely far too many devices to work with and manage. It has become clear that we'll need much higher efficiency user experiences that can allow us to more easily and scalably work with the thousands of devices that will soon be in each of our lives. Enter the conversational interface revolution, combining bots we can literally talk with, gesture to, and even direct with our thoughts, with embedded artificial intelligence, whic...
As DevOps methodologies expand their reach across the enterprise, organizations face the daunting challenge of adapting related cloud strategies to ensure optimal alignment, from managing complexity to ensuring proper governance. How can culture, automation, legacy apps and even budget be reexamined to enable this ongoing shift within the modern software factory? In her Day 2 Keynote at @DevOpsSummit at 21st Cloud Expo, Aruna Ravichandran, VP, DevOps Solutions Marketing, CA Technologies, was jo...
"Storpool does only block-level storage so we do one thing extremely well. The growth in data is what drives the move to software-defined technologies in general and software-defined storage," explained Boyan Ivanov, CEO and co-founder at StorPool, in this SYS-CON.tv interview at 16th Cloud Expo, held June 9-11, 2015, at the Javits Center in New York City.
As Marc Andreessen says software is eating the world. Everything is rapidly moving toward being software-defined – from our phones and cars through our washing machines to the datacenter. However, there are larger challenges when implementing software defined on a larger scale - when building software defined infrastructure. In his session at 16th Cloud Expo, Boyan Ivanov, CEO of StorPool, provided some practical insights on what, how and why when implementing "software-defined" in the datacent...
Blockchain. A day doesn’t seem to go by without seeing articles and discussions about the technology. According to PwC executive Seamus Cushley, approximately $1.4B has been invested in blockchain just last year. In Gartner’s recent hype cycle for emerging technologies, blockchain is approaching the peak. It is considered by Gartner as one of the ‘Key platform-enabling technologies to track.’ While there is a lot of ‘hype vs reality’ discussions going on, there is no arguing that blockchain is b...
Blockchain is a shared, secure record of exchange that establishes trust, accountability and transparency across business networks. Supported by the Linux Foundation's open source, open-standards based Hyperledger Project, Blockchain has the potential to improve regulatory compliance, reduce cost as well as advance trade. Are you curious about how Blockchain is built for business? In her session at 21st Cloud Expo, René Bostic, Technical VP of the IBM Cloud Unit in North America, discussed the b...
You know you need the cloud, but you’re hesitant to simply dump everything at Amazon since you know that not all workloads are suitable for cloud. You know that you want the kind of ease of use and scalability that you get with public cloud, but your applications are architected in a way that makes the public cloud a non-starter. You’re looking at private cloud solutions based on hyperconverged infrastructure, but you’re concerned with the limits inherent in those technologies.
Is advanced scheduling in Kubernetes achievable?Yes, however, how do you properly accommodate every real-life scenario that a Kubernetes user might encounter? How do you leverage advanced scheduling techniques to shape and describe each scenario in easy-to-use rules and configurations? In his session at @DevOpsSummit at 21st Cloud Expo, Oleg Chunikhin, CTO at Kublr, answered these questions and demonstrated techniques for implementing advanced scheduling. For example, using spot instances and co...
The use of containers by developers -- and now increasingly IT operators -- has grown from infatuation to deep and abiding love. But as with any long-term affair, the honeymoon soon leads to needing to live well together ... and maybe even getting some relationship help along the way. And so it goes with container orchestration and automation solutions, which are rapidly emerging as the means to maintain the bliss between rapid container adoption and broad container use among multiple cloud host...
The cloud era has reached the stage where it is no longer a question of whether a company should migrate, but when. Enterprises have embraced the outsourcing of where their various applications are stored and who manages them, saving significant investment along the way. Plus, the cloud has become a defining competitive edge. Companies that fail to successfully adapt risk failure. The media, of course, continues to extol the virtues of the cloud, including how easy it is to get there. Migrating...
Imagine if you will, a retail floor so densely packed with sensors that they can pick up the movements of insects scurrying across a store aisle. Or a component of a piece of factory equipment so well-instrumented that its digital twin provides resolution down to the micrometer.
The need for greater agility and scalability necessitated the digital transformation in the form of following equation: monolithic to microservices to serverless architecture (FaaS). To keep up with the cut-throat competition, the organisations need to update their technology stack to make software development their differentiating factor. Thus microservices architecture emerged as a potential method to provide development teams with greater flexibility and other advantages, such as the abili...
In his keynote at 18th Cloud Expo, Andrew Keys, Co-Founder of ConsenSys Enterprise, provided an overview of the evolution of the Internet and the Database and the future of their combination – the Blockchain. Andrew Keys is Co-Founder of ConsenSys Enterprise. He comes to ConsenSys Enterprise with capital markets, technology and entrepreneurial experience. Previously, he worked for UBS investment bank in equities analysis. Later, he was responsible for the creation and distribution of life settle...
Product connectivity goes hand and hand these days with increased use of personal data. New IoT devices are becoming more personalized than ever before. In his session at 22nd Cloud Expo | DXWorld Expo, Nicolas Fierro, CEO of MIMIR Blockchain Solutions, will discuss how in order to protect your data and privacy, IoT applications need to embrace Blockchain technology for a new level of product security never before seen - or needed.
Leading companies, from the Global Fortune 500 to the smallest companies, are adopting hybrid cloud as the path to business advantage. Hybrid cloud depends on cloud services and on-premises infrastructure working in unison. Successful implementations require new levels of data mobility, enabled by an automated and seamless flow across on-premises and cloud resources. In his general session at 21st Cloud Expo, Greg Tevis, an IBM Storage Software Technical Strategist and Customer Solution Architec...
Nordstrom is transforming the way that they do business and the cloud is the key to enabling speed and hyper personalized customer experiences. In his session at 21st Cloud Expo, Ken Schow, VP of Engineering at Nordstrom, discussed some of the key learnings and common pitfalls of large enterprises moving to the cloud. This includes strategies around choosing a cloud provider(s), architecture, and lessons learned. In addition, he covered some of the best practices for structured team migration an...
In his general session at 21st Cloud Expo, Greg Dumas, Calligo’s Vice President and G.M. of US operations, discussed the new Global Data Protection Regulation and how Calligo can help business stay compliant in digitally globalized world. Greg Dumas is Calligo's Vice President and G.M. of US operations. Calligo is an established service provider that provides an innovative platform for trusted cloud solutions. Calligo’s customers are typically most concerned about GDPR compliance, application p...
Coca-Cola’s Google powered digital signage system lays the groundwork for a more valuable connection between Coke and its customers. Digital signs pair software with high-resolution displays so that a message can be changed instantly based on what the operator wants to communicate or sell. In their Day 3 Keynote at 21st Cloud Expo, Greg Chambers, Global Group Director, Digital Innovation, Coca-Cola, and Vidya Nagarajan, a Senior Product Manager at Google, discussed how from store operations and ...
In his session at 21st Cloud Expo, Raju Shreewastava, founder of Big Data Trunk, provided a fun and simple way to introduce Machine Leaning to anyone and everyone. He solved a machine learning problem and demonstrated an easy way to be able to do machine learning without even coding. Raju Shreewastava is the founder of Big Data Trunk (www.BigDataTrunk.com), a Big Data Training and consulting firm with offices in the United States. He previously led the data warehouse/business intelligence and B...
"IBM is really all in on blockchain. We take a look at sort of the history of blockchain ledger technologies. It started out with bitcoin, Ethereum, and IBM evaluated these particular blockchain technologies and found they were anonymous and permissionless and that many companies were looking for permissioned blockchain," stated René Bostic, Technical VP of the IBM Cloud Unit in North America, in this SYS-CON.tv interview at 21st Cloud Expo, held Oct 31 – Nov 2, 2017, at the Santa Clara Conventi...