DEVOPS WEEKLY ISSUE #280 - 8th May 2016

4 minute read

A pretty wide range of content this week; from in depth posts about how syscalls work to tips on how to embrace a security-minded culture, and from containers and microservices to the evolving role of the systems administrator. Enjoy.

Sponsor

The DevOps Roadtrip is heading to Seattle. Join us in the Space Needle for a half-day of learning & sharing best practices
http://try.victorops.com/devops-roadtrip-seattle

Sponsored

[eBook] 7 Performance Metrics to Release Better Software, Faster

Get real-life examples of application failures and how to avoid them by using 7 key metrics to find problems early on. Also, learn how stuff really works in “the life of a metric” and how to integrate performance metrics into automation tools throughout your application delivery chain.
http://ow.ly/U3BJ7

News

A good description of moving routing logic into the RPC layer in order to solve common problems experienced with a move towards a microservices architecture.
https://blog.buoyant.io/2016/05/04/real-world-microservices-when-services-stop-playing-well-and-start-getting-real/

A well set out series of antipatterns when adopting continuous delivery practices. Includes everything from logging to tips on pipeline design to project planning advice.
http://www.slideshare.net/SkeltonThatcher/continuous-delivery-antipatterns-from-the-wild-matthew-skelton-continuous-lifecycle-london-2016

Lots of people reference the chaos monkey application but it does require a degree of setting up. This post demonstrates packaging chaos monkey in Docker, and then shows how to use it to test your EC2 infrastructure.
https://medium.com/production-ready/chaos-monkey-for-fun-and-profit-87e2f343db31#.v932txkc2

I’ve made similar arguments before, but this post does a great job of reminding folks that most system administrators have been programmers for a while, and similar roles will likely involve more code not less over time
http://pcable.net/2016/05/02/programming/

A little artificial, but a well reasoned argument for why adopting containers is the right approach to building software today, at least from the developers point of view.
http://nickjanetakis.com/blog/save-yourself-from-years-of-turmoil-by-using-docker-today

Containers tend to be used for some degree of isolation, but proper sandboxing requires unprivileged users and more. This post explains the details and demonstrates some potential solutions.
https://blog.jessfraz.com/post/getting-towards-real-sandbox-containers/

If you’ve ever wondered how syscalls work this post is for you. An excellent, detailed but highly readable look at a low level part of Illumos. Good references for anyone wanting more too.
http://zinascii.com/2016/the-illumos-syscall-handler.html

Another of the Craft talks I enjoyed, this one about improving security by helping curate a security conscious culture. Lots of tips for how to avoid the security says No problem.
https://speakerdeck.com/iodboi/building-security-through-culture-craft-conf-budapest-2016

ClusterHQ are conducting a survey into container adoption. Surveys are nearly always more useful the wider the group of people who complete them - container surveys in particular tend to skew towards people using containers in my experience. Please take a moment and fill it it.
https://www.surveymonkey.com/r/5SY6SFQ

Jobs

Close.io is looking for a DevOps / Infrastructure engineer. We’re a profitable fast-growing SaaS startup looking for great people to join our ~15 person team remotely. Come help us improve API performance, tune our databases, tighten up security, setup AWS autoscaling, make deployments faster and safer, scale our Elasticsearch/MySQL/MongoDB/Redis data stores, setup centralized logging, instrument our app with metric collection, set up better monitoring, etc.
https://jobs.lever.co/close.io/38d0c4ac-c3eb-47e9-a49e-4611f96eef8d?lever_source=DevOpsWeekly

FULL STACK ENGINEER - DEVOPS - LEEDS Sky Betting and Gaming (Leeds & Sheffield) are on a mission to find talented individuals with the experience of supporting web applications at scale. You’ll be the driving force for operational excellence within the product teams, working on automated testing and deployment of applications and platforms using Chef. Salary and benefits packages reflect the high standard in our tech teams, with excellent training and development opportunities for all engineers.
http://skybetcareers.com/technical-and-software#R000334

FULL STACK ENGINEER - DEVOPS - SHEFFIELD Come and work with the Sky Betting and Gaming Account team bridging the gap between Operations and Development. You’ll be maintaining and tuning systems that manage payments and customer details. We need people who want to scale systems from 1000’s of requests per second to hundreds of thousands and work closely with development teams so new products seamlessly deploy into production. We scale out NodeJS, Java and php applications using tools such as Chef, Jenkins, NewRelic and Graphite. We cache and store data in Memcache, Redis, kafka, MySQL and Couchbase. This is an opportunity to join us in making Sky Betting and Gaming the biggest and fastest online Gaming and Betting platform in the world.
http://skybetcareers.com/technical-and-software#R000335

Tools

JMESPath is a query language for JSON. It comes with multiple compatible implementations in different languages, all based on a solid specification and test suite. I can see multiple uses for this.
http://jmespath.org/

Knox is another secrets management server with support for key rotation, auditing secrets access and with a desire to be developer friendly.
https://github.com/pinterest/knox

I’m a fan of using Cucumber style tests for infrastructure acceptance. The varnishtest cucumber project provides a set of steps for testing your VCL configurations, some good examples in the README.
https://github.com/nstielau/varnishtest_cucumber

Unik is an interesting attempt to build a unikernel deployment platform. It currently supports C/C++, Go (targeting rumprun) and Java (targeting OSv) and deployments to AWS, vSphere and Virtualbox.
https://github.com/emc-advanced-dev/unik

Many organisations using the same tools adopt different workflows to accomplish similar results. This Chef plugin helps sync local cookbooks with a Chef server, based around a workflow that eschews cookbook versioning.
https://github.com/PagerDuty/pd-sync-chef

The DevOps Roadtrip is heading to Seattle. Join us in the Space Needle for a half-day of learning & sharing best practices
http://try.victorops.com/devops-roadtrip-seattle

Updated: