The document discusses findings from analyzing performance data from 8 million web pages and 1.5 million visits across 50 dimensions. It shows that very fast load times correlate with high bounce rates, and identifies load time thresholds at which bounce rates exceed 50% for different browsers. These "LD50" thresholds are 3.6 seconds for IE, 5.6 seconds for Firefox, and 6.5 seconds for Chrome. The document also analyzes performance trends by country, with Germany, the US, the UK, and Canada each having different average load time thresholds. It concludes by discussing potential areas for future analysis related to user behavior and performance.
(SEC309) Amazon VPC Configuration: When Least Privilege Meets the Penetration...Amazon Web Services
Enterprises trying to deploy infrastructure to the cloud and independent software companies trying to deliver a service have similar problems to solve. They need to know how to create an environment in AWS that enforces least-privilege access between components while also allowing administration and change management. Amazon Elastic Cloud Compute (EC2) and Identity and Access Management (IAM), coupled with services like AWS Security Token Service (STS), offer the necessary building blocks. In this session, we walk through some of the mechanisms available to control access in an Amazon Virtual Private Cloud (VPC). Next, we focus on using IAM and STS to create a least-privilege access model. Finally, we discuss auditing strategies to catch common mistakes and discuss techniques to audit and maintain your infrastructure.
Vik Bhatti (Beamly) - Service Discovery for DevOpsOutlyer
What is service discovery? In the modern world of containers and highly distributed systems, service discovery is mentioned time and time again, but details on how and why it should be implemented are pretty thin. In this talk I will walk through the trials and tribulations of how Beamly has approached the problem, and how the solutions have changed over the past 3 years as the architecture as grown.
Video: https://www.youtube.com/watch?v=-XBONNrb50w
Join DevOps Exchange London here: http://www.meetup.com/DevOps-Exchange-London
Follow DOXLON on twitter http://www.twitter.com/doxlon
The talk from DevOps Days Silicon Valley 2015 conference which describes the signs of having or being a single point of failure expert on your system, and the ways to solve the problem
Microservices and elastic resource pools with Amazon EC2 Container ServiceBoyan Dimitrov
This talk explores a scalable and cost efficient way of deploying and running microservices workloads using quality of service scheduling on top of Amazon EC2 Container service. Running services in a pay as you go fashion will soon be a reality as much as todays on demand compute
(SEC309) Amazon VPC Configuration: When Least Privilege Meets the Penetration...Amazon Web Services
Enterprises trying to deploy infrastructure to the cloud and independent software companies trying to deliver a service have similar problems to solve. They need to know how to create an environment in AWS that enforces least-privilege access between components while also allowing administration and change management. Amazon Elastic Cloud Compute (EC2) and Identity and Access Management (IAM), coupled with services like AWS Security Token Service (STS), offer the necessary building blocks. In this session, we walk through some of the mechanisms available to control access in an Amazon Virtual Private Cloud (VPC). Next, we focus on using IAM and STS to create a least-privilege access model. Finally, we discuss auditing strategies to catch common mistakes and discuss techniques to audit and maintain your infrastructure.
Vik Bhatti (Beamly) - Service Discovery for DevOpsOutlyer
What is service discovery? In the modern world of containers and highly distributed systems, service discovery is mentioned time and time again, but details on how and why it should be implemented are pretty thin. In this talk I will walk through the trials and tribulations of how Beamly has approached the problem, and how the solutions have changed over the past 3 years as the architecture as grown.
Video: https://www.youtube.com/watch?v=-XBONNrb50w
Join DevOps Exchange London here: http://www.meetup.com/DevOps-Exchange-London
Follow DOXLON on twitter http://www.twitter.com/doxlon
The talk from DevOps Days Silicon Valley 2015 conference which describes the signs of having or being a single point of failure expert on your system, and the ways to solve the problem
Microservices and elastic resource pools with Amazon EC2 Container ServiceBoyan Dimitrov
This talk explores a scalable and cost efficient way of deploying and running microservices workloads using quality of service scheduling on top of Amazon EC2 Container service. Running services in a pay as you go fashion will soon be a reality as much as todays on demand compute
(Talk given at Continuous Lifecycle London 2016)
Continuous Delivery techniques and practices are often misunderstood. This session will explore some Continuous Delivery anti-patterns based on work 'in the wild' with a wide range of organisations across different industry sectors:
- Believing that "Continuous Delivery is not for us"
- Ignoring the database
- Thinking that a deployment pipeline is just a series of chained jobs in Jenkins
- Not funding the build/test/deployment capability properly
- No effective logging or application metrics
By avoiding these pitfalls, we can increase the effectiveness of our software delivery efforts.
Who Does What? Mapping Cloud Foundry Activities and Entitlements to IT RolesVMware Tanzu
SpringOne Platform 2016
Speaker: Cornelia Davis; Sr. Director of Technology, Pivotal.
While “cf push” is the center of it all, there are many more things that various individuals can do with the Cloud Foundry platform. They can monitor, scale and upgrade those deployed apps. And also deploy, monitor, scale and upgrade the platform itself. Further, to operationalize the platform in an enterprise there are quotas, security groups, route services, environment variable groups and many other “knobs” that may also be tuned, and there are various roles and permission structures to govern these. In this session Cornelia will take a holistic view of the Cloud Foundry “control plane” and map the key functions to IT roles (perhaps with some redefinition), and she’ll show which entitlements allow which configurations. Ultimately the goal is to understand how Cloud Foundry can be effectively used to optimize the development and operations (Devops) in your organization. Participants will leave with a concrete framework for transforming current IT practices, roles and responsibilities using the Cloud Foundry platform.
Identifying and fixing issues in new code before deploying it to production is important for every software development cycle. However, relying on traditional testing methods in the age of Internet-scale data driven problems may prove to be incomplete. Identifying and fixing the issues in production quickly is crucial, but it requires insight into usage patterns and trends across the whole architecture and application logic. In this talk I touch on inefficiencies of some of the most common testing methods, provide real world examples of discovering odd edge cases with monitoring and offer recommendations on top-down metric instrumentation to help DevOps organizations with identifying and acting on business-effecting problems.
A talk about lessons learned from building developer and sysadmin-facing tools at Puppet and Docker. It’s applicable to open source tools for broader use and the tools you and your teams develop inside your organization. Building tools, both external and internal, is hard. Getting people to adopt those tools is even harder. These are ideas to make it easier.
Inspec, or how to translate compliance spreadsheets into codeMichael Goetz
InSpec allows you to examine any node with controls that can written in simple form and then executed in an automated fashion as part of your software development process. We'll talk about the basic concepts of InSpec, how to write controls and how to use the reported output to take your compliance spreadsheets into a automated development world.
Monitorama 2015 talk by Brendan Gregg, Netflix. With our large and ever-changing cloud environment, it can be vital to debug instance-level performance quickly. There are many instance monitoring solutions, but few come close to meeting our requirements, so we've been building our own and open sourcing them. In this talk, I will discuss our real-world requirements for instance-level analysis and monitoring: not just the metrics and features we desire, but the methodologies we'd like to apply. I will also cover the new and novel solutions we have been developing ourselves to meet these needs and desires, which include use of advanced Linux performance technologies (eg, ftrace, perf_events), and on-demand self-service analysis (Vector).
Talk for SCaLE13x. Video: https://www.youtube.com/watch?v=_Ik8oiQvWgo . Profiling can show what your Linux kernel and appliacations are doing in detail, across all software stack layers. This talk shows how we are using Linux perf_events (aka "perf") and flame graphs at Netflix to understand CPU usage in detail, to optimize our cloud usage, solve performance issues, and identify regressions. This will be more than just an intro: profiling difficult targets, including Java and Node.js, will be covered, which includes ways to resolve JITed symbols and broken stacks. Included are the easy examples, the hard, and the cutting edge.
There are two common tenets of operations: "hell is other people's software," and "better software is produced by those forced to operate it." In this session I'll take a fly-by-tour of two pieces of software that were built from the ground up for operability from the hard-earned teachings of their inoperable predecessors: a distributed datastore replacing PostgreSQL, and a message queue replacing RabbitMQ.
We'll discuss specific design aspects that increase resiliency in the event of failure and observability at all times.
Speed is Essential for a Great Web ExperienceDigicure ApS
Andy Davies' slides fra Digicures web performance seminar på Bella Sky, 24. oktober 2012.
Den engelske ekspert og freelance konsulent, Andy Davies, har været bidt af emnet om god web performance siden 90'erne. Andy hjælper virksomheder med at måle, analysere og forbedre performance og stabiliteten af deres websites.
Selvom hastighed er vigtig for en god online oplevelse, bliver det ofte overset. Andy Davies vil fortælle hvordan hastighed påvirker brugeroplevelsen på et website og afdække hvordan hastighed kan måles og analyseres.
You know that performance is crucial to your company's success, but do the people in the corner office know this? You need to get the message across using the language they speak and targeting the goals they care about.
This session -- presented by Strangeloop president Joshua Bixby at the 2011 Web Performance Summit -- summarizes the benefits of a faster website or web app, then delves into a series of how-tos for creating a business case for web performance in your organization.
(Talk given at Continuous Lifecycle London 2016)
Continuous Delivery techniques and practices are often misunderstood. This session will explore some Continuous Delivery anti-patterns based on work 'in the wild' with a wide range of organisations across different industry sectors:
- Believing that "Continuous Delivery is not for us"
- Ignoring the database
- Thinking that a deployment pipeline is just a series of chained jobs in Jenkins
- Not funding the build/test/deployment capability properly
- No effective logging or application metrics
By avoiding these pitfalls, we can increase the effectiveness of our software delivery efforts.
Who Does What? Mapping Cloud Foundry Activities and Entitlements to IT RolesVMware Tanzu
SpringOne Platform 2016
Speaker: Cornelia Davis; Sr. Director of Technology, Pivotal.
While “cf push” is the center of it all, there are many more things that various individuals can do with the Cloud Foundry platform. They can monitor, scale and upgrade those deployed apps. And also deploy, monitor, scale and upgrade the platform itself. Further, to operationalize the platform in an enterprise there are quotas, security groups, route services, environment variable groups and many other “knobs” that may also be tuned, and there are various roles and permission structures to govern these. In this session Cornelia will take a holistic view of the Cloud Foundry “control plane” and map the key functions to IT roles (perhaps with some redefinition), and she’ll show which entitlements allow which configurations. Ultimately the goal is to understand how Cloud Foundry can be effectively used to optimize the development and operations (Devops) in your organization. Participants will leave with a concrete framework for transforming current IT practices, roles and responsibilities using the Cloud Foundry platform.
Identifying and fixing issues in new code before deploying it to production is important for every software development cycle. However, relying on traditional testing methods in the age of Internet-scale data driven problems may prove to be incomplete. Identifying and fixing the issues in production quickly is crucial, but it requires insight into usage patterns and trends across the whole architecture and application logic. In this talk I touch on inefficiencies of some of the most common testing methods, provide real world examples of discovering odd edge cases with monitoring and offer recommendations on top-down metric instrumentation to help DevOps organizations with identifying and acting on business-effecting problems.
A talk about lessons learned from building developer and sysadmin-facing tools at Puppet and Docker. It’s applicable to open source tools for broader use and the tools you and your teams develop inside your organization. Building tools, both external and internal, is hard. Getting people to adopt those tools is even harder. These are ideas to make it easier.
Inspec, or how to translate compliance spreadsheets into codeMichael Goetz
InSpec allows you to examine any node with controls that can written in simple form and then executed in an automated fashion as part of your software development process. We'll talk about the basic concepts of InSpec, how to write controls and how to use the reported output to take your compliance spreadsheets into a automated development world.
Monitorama 2015 talk by Brendan Gregg, Netflix. With our large and ever-changing cloud environment, it can be vital to debug instance-level performance quickly. There are many instance monitoring solutions, but few come close to meeting our requirements, so we've been building our own and open sourcing them. In this talk, I will discuss our real-world requirements for instance-level analysis and monitoring: not just the metrics and features we desire, but the methodologies we'd like to apply. I will also cover the new and novel solutions we have been developing ourselves to meet these needs and desires, which include use of advanced Linux performance technologies (eg, ftrace, perf_events), and on-demand self-service analysis (Vector).
Talk for SCaLE13x. Video: https://www.youtube.com/watch?v=_Ik8oiQvWgo . Profiling can show what your Linux kernel and appliacations are doing in detail, across all software stack layers. This talk shows how we are using Linux perf_events (aka "perf") and flame graphs at Netflix to understand CPU usage in detail, to optimize our cloud usage, solve performance issues, and identify regressions. This will be more than just an intro: profiling difficult targets, including Java and Node.js, will be covered, which includes ways to resolve JITed symbols and broken stacks. Included are the easy examples, the hard, and the cutting edge.
There are two common tenets of operations: "hell is other people's software," and "better software is produced by those forced to operate it." In this session I'll take a fly-by-tour of two pieces of software that were built from the ground up for operability from the hard-earned teachings of their inoperable predecessors: a distributed datastore replacing PostgreSQL, and a message queue replacing RabbitMQ.
We'll discuss specific design aspects that increase resiliency in the event of failure and observability at all times.
Speed is Essential for a Great Web ExperienceDigicure ApS
Andy Davies' slides fra Digicures web performance seminar på Bella Sky, 24. oktober 2012.
Den engelske ekspert og freelance konsulent, Andy Davies, har været bidt af emnet om god web performance siden 90'erne. Andy hjælper virksomheder med at måle, analysere og forbedre performance og stabiliteten af deres websites.
Selvom hastighed er vigtig for en god online oplevelse, bliver det ofte overset. Andy Davies vil fortælle hvordan hastighed påvirker brugeroplevelsen på et website og afdække hvordan hastighed kan måles og analyseres.
You know that performance is crucial to your company's success, but do the people in the corner office know this? You need to get the message across using the language they speak and targeting the goals they care about.
This session -- presented by Strangeloop president Joshua Bixby at the 2011 Web Performance Summit -- summarizes the benefits of a faster website or web app, then delves into a series of how-tos for creating a business case for web performance in your organization.
London Web Performance Meetup: Performance for mortal companiesStrangeloop
You're probably familiar with the well-known performance success stories from companies like Amazon, Google, Microsoft and Shopzilla. But how relevant are these megasites to "mortal companies" that don't make billions of dollars per year or have teams of in-house performance engineers to do their bidding?
Strangeloop president Joshua Bixby walks through case studies of Strangeloop customers like AutoAnything.com and Artbeads.com to show how mortal companies have improved performance and achieved measurable success, including:
· Increased revenue by 13%
· Increased cart size by 6%
· Increased conversions by 9%
Joshua offers practical tips for successfully evangelizing performance within your organization. He also gives a snapshot of the current performance landscape in North America, as well as a sense of where the industry is headed.
This plugin implements a flyweight design pattern for jQuery custom select controls. By only holding one visible control element in memory at one time, it aims to reduce the number of DOM accesses and event binding operations. This can result in significantly reduced page render times on larger forms. It borrows accessibility design patterns from the Filament Group's jQuery UI Selectmenu keyboard navigable ARIA plugin, currently included with JQuery-UI. It requires jQuery 1.4+ and jQuery UI 1.8.7+ Core to run.
MW2010: Nate Solas, Hiding our Collections in Plain Site: Interface Strategie...museums and the web
A presentation from Museums and the Web 2010.
A recent redesign of the collection search interface for an on-line art education tool (ArtsConnectEd.org) has provided an opportunity to compare usage patterns between the two versions. In this paper I first survey current search interface design patterns, then discuss the new interface, the log cleanup and analysis, and finally present evidence-based recommendations that may be applied to the general problem of presenting large collections on-line.
see http://www.archimuse.com/mw2010/abstracts/prg_335002302.html
Slide deck for a presentation at OSCON 2011 about why Netflix uses web technology for TV user interfaces and how we maximize performance for a broad range of devices.
Speaker: Joseph Rea, Engineer, Confluent
Joseph will talk about how to visualize topics in Apache Kafka®, the difference between a stream and table in KSQL and his lessons learned on tackling this technical challenge with millions of Kafka messages consumed per second. With such functionalities, users can understand their data easier and in a highly performant and scalable way. This talk covers understanding web workers as they relate to webpack, web socket management, debugging browser performance and the future of the applications that can now be built.
Joseph Rea started engineering with the LAMP stack building custom e-commerce checkouts, ERP systems and enterprise water/sewer billing software. He worked at Yahoo as a front end engineer in the media org before doing Android and iOS development for the video SDK. He also worked at LifeLock to build an application that updated PII on various service sites. He likes turtles. He currently works at Confluent building so much UI. He blogs at https://cnfl.io/blog-joseph-rea.
RUM for Breakfast - distilling insights from the noiseBuddy Brewer
From Velocity 2012 in Santa Clara, CA. Buddy Brewer, Philip Tellis, and Carlos Bueno talk about real user measurement collection, analysis, and insights.
The 3.5s Dash for Attention and Other Stuff We Found in RUM
1. Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 1
2. The 3.5s dash for attention and other stuff we
found in RUM
Buddy Brewer, Philip Tellis
Velocity EU 2012 / 2012-10-03
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 2
3. Jean-Antoine Nollet
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 3
4. Mr. Souders
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 4
6. Never underestimate the bandwidth of a mail horse
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 6
7. World-wide bandwidth
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 7
8. What do users expect?
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 8
9. Gold or Bronze?
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 9
10. What delighted users a few years ago is now an expected
baseline, the absence of which will frustrate.
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 10
11. Should you delight?
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 11
12. Or frustrate?
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 12
13. Delighting changes the baseline
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 13
14. Delighting changes the baseline
(Frustrating does too, but we don’t want to go there)
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 13
15. So where’s the web today?
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 14
16. Log-Normal Distribution
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 15
17. Log-Normal Distribution
The logarithm of the x-axis follows a Normal distribution
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 15
18. Log-Normal Distribution
Performance data does not always follow a "pure" Log-Normal
distribution
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 16
19. Look at the entire spread
...
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 17
20. Look at the entire spread
which often approaches an infinite width
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 17
21. • 0.8% of hits are fake/abusive
• 0.2-0.5% of hits are from a stale cache
• 0.1% of hits are absurd
• Timestamps in the future (or past depending on how you
interpret it)
• Bots ignore robots.txt across domains
• "Interesting" caches/copies
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 18
22. Even with beacons, you need to sanitize your input
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 19
23. Once you get past the cruft, you can really measure users
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 20
29. bounce rate vs. front end time
80.00%
60.00%
40.00%
20.00%
0%
0.5 2 3.5 5 6.5 8 9.5 11 12.5 14 15.5 17 18.5 20 21.5 23 24.5 26 27.5 29
Wednesday, October 3, 12
30. is my web site performance toxic to my
users?
http://www.flickr.com/photos/21560098@N06/3796822070
Wednesday, October 3, 12
31. LD50 - when do half the users bounce?
http://www.flickr.com/photos/thecosmopolitan/6117530924
Wednesday, October 3, 12
32. Bounce rate >=50%
Back end time 1.7 sec
DOM Loading 1.8 sec
DOM Interactive 2.75 sec
Front end time 3.5 sec
DOM Complete 4.75 sec
Load event 5.5 sec
Wednesday, October 3, 12
37. Future directions
What is the LD50 for your site?
Other bounce rates? 40%? 30%?
Other variables? (critical content
visible, etc)
Other behaviors? Conversions,
revenue, pages per session, actions,
when do people make tea?
Wednesday, October 3, 12
38. Questions?
Buddy Brewer @bbrewer
Philip Tellis @bluesmoon
Wednesday, October 3, 12
39. Credits
• Usain Bolt – Sum_of_Mark on flickr
• Douglas Mail Carrier – Svadilfari on flickr
• Angel Delight – Auntie P on flickr
• Frustrated – Kevin Lawver on flickr
• Jean-Antoine Nollet – (Public Domain) Wikipedia
• 100 metre dash – on NY Times
Velocity EU 2012 / 2012-10-03 The 3.5s dash for attention and other stuff we found in RUM 21