DEVOPS WEEKLY ISSUE #202 - 16th November 2014
With all the renewed interest in Windows automation recently, I just wanted to say congratulations to Rob Reynolds on the success of the Chocolatey Kickstarter. https://chocolatey.org/ is apt-get or yum for Windows and a great example of the open source community making Windows better.
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
We’re a week away from the New Directions in Operating Systems conference, so this post is particularly relevant. It’s a good introduction to unikernels and why we’ll likely see lots of change in the operating system space in the next year or so.
https://medium.com/@darrenrush/after-docker-unikernels-and-immutable-infrastructure-93d5a91c849e
A look under the hood of how Netflix get maximum performance from running on EC2. Lots of information and tips on everything from choosing instance types to kernel tuning and managing costs.
http://www.slideshare.net/brendangregg/performance-tuning-ec2-instances
A good post questioning whether you need a large framework for system testing. I’ve definitely seen the issues described here.
http://codedetective.blogspot.co.uk/2014/09/do-you-really-need-test-frameworks.html
A video and slide deck telling the Instagram scaling story, from a single server to AWS and then to Facebook. Nicely focuses on lots of human stories as well as the organisation grew.
https://opbeat.com/blog/posts/how-we-scaled-instagram-mike-krieger/
A great looking presentation entitled “The Hidden Drivers of Devops and Continuous Delivery”, lots of good points about culture, empathy, management and human error.
http://slidr.io/roidrage/the-hidden-drivers-of-devops-and-continuous-delivery#46
A nice post relating culture, automation and measurement to the desire to operationalize our systems.
https://devcentral.f5.com/articles/just-what-does-operationalize-mean-anyway
Increasing the frequency of deployments means getting over the fear of shipping code to production. This post describes one organisation’s experiences with using architecture choices, monitoring and testing to get over that fear.
http://tech.ticketmaster.com/2014/11/14/fear-and-paranoia-in-a-devops-world/
We can all probably agree on the value of automated testing, but moving an existing product or team to new ways of working is difficult. So real world stories like the following are interesting.
http://servicemanagement360.com/2014/11/13/building-culture-test-automation-journey-devops/
Jobs
Ooyala is hiring Network Engineers to help design and scale the network infrastructure in our private OpenStack and cloud (Azure and AWS) environments. Automate our routing and switching (http://bit.ly/10jgk37)) infrastructure with Chef and other configuration management tooling to allow us to scale quickly. The solutions and tooling you build will be used to power video on some of your favorite websites.
https://hire.jobvite.com/j?aj=o0auZfwn&s=DevOps_Weekly.
MediaMath is looking for an SRE to help build and manage our infrastructure and data centers across four continents. Our SREs are front-and-center in the effort to design, build, and keep our distributed services fast and reliable, 100% of the time. If the stability and performance of real-time, global systems sounds like an exciting challenge, join our team.
https://careers-mediamath.icims.com/jobs/1538/site-reliability-engineer—techops/job
Tools
Bosun is an all-in-one monitoring system backed by OpenTSDB. It provides alerting, time series data collection and visualisation as well as metric collection. It also treats both Linux and Windows as first class citizens.
http://bosun.org/
Octopush is another tool for using Jenkins for deployment automation. It provides a dedicated user interface and API, as well as integration with GitHub for authentication.
https://github.com/olx-inc/octopush
http://olx-inc.github.io/octopush/