DEVOPS WEEKLY ISSUE #171 - 13th April 2014

This issue of Devops Weekly is dedicated to everyone who spent a good portion of their week resolving issues arising from the Heartbleed OpenSSL bug. I’m guessing that’s quite a few of you.

Sponsor

Devops Weekly is sponsored by Brightbox Cloud - serious UK-based cloud infrastructure from only 1.5p per hour (£10.95/month)

Start your £20 free trial now: http://brightbox.com/devopsweekly

News

Lots of readers of this newsletter will hopefully be having a well earned rest this weekend after the announcement on Monday of a serious bug in OpenSSL that has incredibly widespread implications. The bug, named Heartbleed, affects everyone. If you’re not already on top of this here’s a few useful posts.
http://heartbleed.com/
http://www.craigkerstiens.com/2014/04/08/What-you-need-to-know-about-april-7-and-the-web/
http://theagileadmin.com/2014/04/08/gauntlt-test-for-heartbleed/

A great presentation deck about what continuous delivery means to the people involved. Focuses on the difficulty of changing medium sized teams and what happens to the developer, tester and systems administrator roles.
http://www.slideshare.net/TomasRiha/continuous-delivery-its-not-about-the-technology-its-about-the-people-pipelineconf

A solid argument that release testing, as a separate activity at least, is a bad idea when looking to take advantage of continuous delivery. Lots of good references too if you’re making this case at work.
http://www.alwaysagileconsulting.com/organisation-antipattern-release-testing/

CoreOS is starting to get very interesting, with the addition of higher levels tools like fleet which give you an idea about the ambition of the project. This presentation does a good job of introducing some of the new functionality.
http://www.slideshare.net/geekle/coreos-control-your-fleet

A good run through of some of the disadvantages or at least challenges when building out a system based on many small service. The demand for, and interest in, operations experience definitely feels like it goes hand-in-hand with this pattern.
http://highscalability.com/blog/2014/4/8/microservices-not-a-free-lunch.html

Short of events to go along to this year that have a devops flavour? Here’s a good round-up of events from local devopsdays to the larger Velocity and LISA.
http://www.slideshare.net/dlutzy/2014-dev-ops-conferences

What is high availability and how do you achieve it? This is always fiddly in my experience, at least for software that’s not distributed by default. This presentation from the recent LOADays conference has some good examples.
http://www.slideshare.net/roidelapluie/loadays-linuxha

Jobs

Kickstarter is looking for an Operations Engineer to work on our infrastructure’s reliability, resiliency, and performance. You’re comfortable with EC2 with Chef, RDS with MySQL, and Ruby on Rails, and handy in Redis, Elasticsearch, Redshift, and CloudFront. If you have experience running production web applications, an interest in creative projects, and get excited about working on internal tools that accelerate development, simplify operations, and enable all employees to contribute to the product, we’d love to talk.
https://www.kickstarter.com/jobs/ops-engineer

Events

The 28th LISA conference is looking for talk submissions in the areas of monitoring, systems engineering, culture, security, and devops. We’ve got an audience that spans industries, and we’d like to make the conference even better with your submission. Talk submissions are being accepted until Friday, 18 April 2014. For those interested, we are also accepting posters and papers too.

www.usenix.org/lisa14

Devopsdays Silicon Valley is scheduled for the 27th and 28th of June and currently has a call for proposals open until the 2nd of May. The page below has a good list of suggested topics to get you thinking.
http://www.devopsdays.org/events/2014-siliconvalley/propose/

Tools

I’ve often used software testing tools for monitoring purposes, but one of the challenges is extracting the relevant information from a test run. This post and set of tools does a nice job of that for Python, as well as providing a good dashboard for raising the visibility of test runs.
http://open.blogs.nytimes.com/2014/04/08/improving-the-user-experience-of-automated-integration-testing/
http://developers.nytimes.com/opensource/sneeze/

An interesting tool that connects MySQL with statsd, simply write SQL queries and provide a metrics name and the script will do the rest.
https://github.com/banyek/Mambo-collector

Envirius is described as a universal virtual environments manager. Like virtualenv in python, rvm in ruby, nvm for node, etc. but without needing a separate tool for every language or framework.
https://github.com/ekalinin/envirius