Welcome!

Microservices Expo Authors: Liz McMillan, Derek Weeks, Elizabeth White, PagerDuty Blog, Pat Romanski

Related Topics: Microservices Expo, Java IoT, Linux Containers, Agile Computing, Cloud Security, @BigDataExpo

Microservices Expo: Article

Twitter’s Story: How Homegrown Load Testing Tools Can Misfire

If Twitter built their own tool to perform their own load test, why did the selfie cause their site to crash?

The selfie that changed the world, or at least Twitter, has been in the news for the past month. On March 2, 2014, the infamous Oscar selfie of Ellen and her celebrity friends broke a record of 2 million retweets before midnight the same night. That record was previously set by President Barack Obama, hugging first lady Michelle Obama after his 2012 re-election.

The selfie caused Twitter to crash for more than 20 minutes, also breaking the record for the longest crash of the social media site. Twitter was infamous for crashing in its early days (anyone remember "Fail Whale?"), so it's no wonder the social media giant worked extra hard to completely prepare their website infrastructure before going public in November 2013. This included building their own load testing tool, Iago, in 2012.

If they built their own tool to perform their own load test, why did the selfie cause their site to crash? The Oscar selfie crash is a perfect example of what companies can easily overlook. Twitter did not test their users properly and their homegrown tool clearly doesn't solve all of their problems... their servers still crash.

What is Iago and why did Twitter decide to make it? And what does that have to do with your decision to use homegrown tools versus vendor load testing tools? Don't worry, we will tell you.

Twitter's Homegrown Load Testing Tool: Iago
Iago was created in June of 2012 by Twitter's internal engineering team. According to Twitter, Iago is a load generator created to help the social media site test services before they encounter production traffic. Chris Aniszczyk, Head of Open Source at Twitter, said, "There are many load generators available in the open source and commercial software worlds, but Iago provided us with capabilities that are uniquely suited for Twitter's environment and the precise degree to which we need to test our services."

Basically their homegrown tool was completely customized for their platform alone - a very attractive aspect of developing your own tool.

The three attributes Twitter focused on in creating Iago were:

  1. High performance: Iago was designed to generate traffic in a precise and predictable way, to minimize variance between test runs and allow comparisons to be made between development iterations.
  2. Multi-protocol: Modeling a system as complex as Twitter can be difficult, but it's made easier by decomposing it into component services. Once decomposed, each piece can be tested in isolation; which requires the load generator to speak each service's protocol. Twitter has in excess of 100 such services, and Iago tests most of them using built-in support for the internal protocols Twitter uses.
  3. Extensible: Iago is designed for engineers. It assumes the person building the system will also be interested in validating performance. As such, the tool is designed from the ground up to be extensible - making it easy to generate new traffic types, over new protocols and with individualized traffic sources.

Why Twitter Couldn't Handle Ellen's Selfie
If we were to do the math, Iago was up and running for nearly two years before the Oscar selfie. What happened to their load testing tool?

There were two main reasons why Twitter crashed. First, the tweet Ellen posted was a picture. On Twitter, a tweet accounts for only 260 bytes of data while a picture on Twitter accounts for 33KB of data, almost 130 times as much as a tweet. Second, Twitter's distributed server system was already at max capacity so the load taken on by the website couldn't be distributed to any nearby servers.

Twitter made one major mistake contributing to the crash back in March: they didn't anticipate and replicate real user activity. Most likely Iago wasn't instructed to generate a load based on a picture being retweeted millions of times, thus Twitter didn't know what to expect when the Oscars rolled around.

Homegrown Tool vs. Vendor Load Testing Solutions
Twitter was looking for a DIY homegrown solution because of their unique platform, and while most load testing tools seek to accomplish the same goal, there are always differences between tools. Here are some of the differences we see between homegrown and vendor-provided load testing tools.

Homegrown

Vendor

1. You can customize it - A homegrown tool, created completely from scratch, means you can build exactly what you think you need. But you don't get the benefit of lots of other people's experiences - so what you think you need may not turn out to be what you actually need.

2. Homegrown tools are free like a puppy is free - We have mentioned this analogy before in another post, and it needs to be brought up again. Homegrown tools aren't really free. You have to pay a team to keep them up and running.

3. What if someone leaves? - A couple of developers are experts on the tool they built, but what happens when they leave? Who is going to be your expert? If the new person isn't properly trained and if the code isn't well documented, then modifications to the code become extremely difficult.

1. Your vendor makes it extensible - A vendor that supports lots of users is constantly adding features and capabilities to support needs you may not have realized you have yet. With extensible APIs and SDKs, you still have the ability to add integrations and the protocol support your application demands.

2. You pay for something that works - With homegrown tools there is no guarantee that it will work 100% of the time. By paying a vendor to use their tool, you have assurances that your load testing tool won't let you down when you need it most.

3. Support Team - Vendors dedicate people to work on any issues and keep you worry-free. Their team of experts is available to make sure all of your questions and concerns are addressed.

Homegrown Tools Are Not for Everyone
Twitter's tool has done well to help them improve the performance of the service, but continued service disruptions show that it isn't perfect. A homegrown tool isn't always the best solution, and most of the time, companies do not have the resources handy to completely build their load testing tool from scratch. If you are interested in learning more about vendor options, give us a call. We're more than happy to help.

More Stories By Tim Hinds

Tim Hinds is the Product Marketing Manager for NeoLoad at Neotys. He has a background in Agile software development, Scrum, Kanban, Continuous Integration, Continuous Delivery, and Continuous Testing practices.

Previously, Tim was Product Marketing Manager at AccuRev, a company acquired by Micro Focus, where he worked with software configuration management, issue tracking, Agile project management, continuous integration, workflow automation, and distributed version control systems.

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.


@MicroservicesExpo Stories
Microservices are a very exciting architectural approach that many organizations are looking to as a way to accelerate innovation. Microservices promise to allow teams to move away from monolithic "ball of mud" systems, but the reality is that, in the vast majority of organizations, different projects and technologies will continue to be developed at different speeds. How to handle the dependencies between these disparate systems with different iteration cycles? Consider the "canoncial problem" ...
After more than five years of DevOps, definitions are evolving, boundaries are expanding, ‘unicorns’ are no longer rare, enterprises are on board, and pundits are moving on. Can we now look at an evolution of DevOps? Should we? Is the foundation of DevOps ‘done’, or is there still too much left to do? What is mature, and what is still missing? What does the next 5 years of DevOps look like? In this Power Panel at DevOps Summit, moderated by DevOps Summit Conference Chair Andi Mann, panelists l...
When building DevOps or continuous delivery practices you can learn a great deal from others. What choices did they make, what practices did they put in place, and how did they connect the dots? At Sonatype, we pulled together a set of 21 reference architectures for folks building continuous delivery and DevOps practices using Docker. Why? After 3,000 DevOps professionals attended our webinar on "Continuous Integration using Docker" discussing just one reference architecture example, we recogn...
Hardware virtualization and cloud computing allowed us to increase resource utilization and increase our flexibility to respond to business demand. Docker Containers are the next quantum leap - Are they?! Databases always represented an additional set of challenges unique to running workloads requiring a maximum of I/O, network, CPU resources combined with data locality.
Thanks to Docker and the DevOps revolution, microservices have emerged as the new way to build and deploy applications — and there are plenty of great reasons to embrace the microservices trend. If you are going to adopt microservices, you also have to understand that microservice architectures have many moving parts. When it comes to incident management, this presents an important difference between microservices and monolithic architectures. More moving parts mean more complexity to monitor an...
In recent years, containers have taken the world by storm. Companies of all sizes and industries have realized the massive benefits of containers, such as unprecedented mobility, higher hardware utilization, and increased flexibility and agility; however, many containers today are non-persistent. Containers without persistence miss out on many benefits, and in many cases simply pass the responsibility of persistence onto other infrastructure, adding additional complexity.
Docker containers have brought great opportunities to shorten the deployment process through continuous integration and the delivery of applications and microservices. This applies equally to enterprise data centers as well as the cloud. In his session at 20th Cloud Expo, Jari Kolehmainen, founder and CTO of Kontena, will discuss solutions and benefits of a deeply integrated deployment pipeline using technologies such as container management platforms, Docker containers, and the drone.io Cl tool...
The rise of containers and microservices has skyrocketed the rate at which new applications are moved into production environments today. While developers have been deploying containers to speed up the development processes for some time, there still remain challenges with running microservices efficiently. Most existing IT monitoring tools don’t actually maintain visibility into the containers that make up microservices. As those container applications move into production, some IT operations t...
In 2014, Amazon announced a new form of compute called Lambda. We didn't know it at the time, but this represented a fundamental shift in what we expect from cloud computing. Now, all of the major cloud computing vendors want to take part in this disruptive technology. In his session at 20th Cloud Expo, John Jelinek IV, a web developer at Linux Academy, will discuss why major players like AWS, Microsoft Azure, IBM Bluemix, and Google Cloud Platform are all trying to sidestep VMs and containers...
As Enterprise business moves from Monoliths to Microservices, adoption and successful implementations of Microservices become more evident. The goal of Microservices is to improve software delivery speed and increase system safety as scale increases. Documenting hurdles and problems for the use of Microservices will help consultants, architects and specialists to avoid repeating the same mistakes and learn how and when to use (or not use) Microservices at the enterprise level. The circumstance w...
SYS-CON Events announced today that CA Technologies has been named “Platinum Sponsor” of SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY, and the 21st International Cloud Expo®, which will take place October 31-November 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA. CA Technologies helps customers succeed in a future where every business – from apparel to energy – is being rewritten by software. From ...
With 10 simultaneous tracks, keynotes, general sessions and targeted breakout classes, Cloud Expo and @ThingsExpo are two of the most important technology events of the year. Since its launch over eight years ago, Cloud Expo and @ThingsExpo have presented a rock star faculty as well as showcased hundreds of sponsors and exhibitors! In this blog post, I provide 7 tips on how, as part of our world-class faculty, you can deliver one of the most popular sessions at our events. But before reading the...
This week's news brings us further reminders that if you're betting on cloud, you're headed in the right direction. The cloud is growing seven times faster than the rest of IT, according to IDC, with a 25% spending increase just from 2016 to 2017. SaaS still leads the pack, with an estimated two-thirds of public cloud spending going that way. Large enterprises, with more than 1,000 employees, are predicted to account for more than half of cloud spending and have the fastest annual growth rate.
Microservices (μServices) are a fascinating evolution of the Distributed Object Computing (DOC) paradigm. Initial design of DOC attempted to solve the problem of simplifying developing complex distributed applications by applying object-oriented design principles to disparate components operating across networked infrastructure. In this model, DOC “hid” the complexity of making this work from the developer regardless of the deployment architecture through the use of complex frameworks, such as C...
@DevOpsSummit at Cloud taking place June 6-8, 2017, at Javits Center, New York City, is co-located with the 20th International Cloud Expo and will feature technical sessions from a rock star conference faculty and the leading industry players in the world. The widespread success of cloud computing is driving the DevOps revolution in enterprise IT. Now as never before, development teams must communicate and collaborate in a dynamic, 24/7/365 environment. There is no time to wait for long developm...
DevOps and microservices are permeating software engineering teams broadly, whether these teams are in pure software shops but happen to run a business, such Uber and Airbnb, or in companies that rely heavily on software to run more traditional business, such as financial firms or high-end manufacturers. Microservices and DevOps have created software development and therefore business speed and agility benefits, but they have also created problems; specifically, they have created software securi...
In today's enterprise, digital transformation represents organizational change even more so than technology change, as customer preferences and behavior drive end-to-end transformation across lines of business as well as IT. To capitalize on the ubiquitous disruption driving this transformation, companies must be able to innovate at an increasingly rapid pace. Traditional approaches for driving innovation are now woefully inadequate for keeping up with the breadth of disruption and change facing...
SYS-CON Events announced today that Outlyer, a monitoring service for DevOps and operations teams, has been named “Bronze Sponsor” of SYS-CON's 20th International Cloud Expo®, which will take place on June 6-8, 2017, at the Javits Center in New York City, NY. Outlyer is a monitoring service for DevOps and Operations teams running Cloud, SaaS, Microservices and IoT deployments. Designed for today's dynamic environments that need beyond cloud-scale monitoring, we make monitoring effortless so you...
Cloud Expo, Inc. has announced today that Andi Mann and Aruna Ravichandran have been named Co-Chairs of @DevOpsSummit at Cloud Expo 2017. The @DevOpsSummit at Cloud Expo New York will take place on June 6-8, 2017, at the Javits Center in New York City, New York, and @DevOpsSummit at Cloud Expo Silicon Valley will take place Oct. 31-Nov. 2, 2017, at the Santa Clara Convention Center in Santa Clara, CA.
All clouds are not equal. To succeed in a DevOps context, organizations should plan to develop/deploy apps across a choice of on-premise and public clouds simultaneously depending on the business needs. This is where the concept of the Lean Cloud comes in - resting on the idea that you often need to relocate your app modules over their life cycles for both innovation and operational efficiency in the cloud. In his session at @DevOpsSummit at19th Cloud Expo, Valentin (Val) Bercovici, CTO of Soli...