3 minute read

Devops weekly now has more than 10,000 subscribers. Given that when I started this newsletter 3 years ago probably only hundreds of people had even heard the word that’s pretty amazing to me. Thankyou to everyone who has written anything I’ve featured, that keeps me interested enough to keep curating this newsletter.


Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…


An excellent post about building a testing culture in an organisation, in this case Airbnb. Honesty about problems and lots of good advice for addressing them.

CloudFormation templates can become unwieldy quickly so the testing workflow this post suggests should be useful. It combines guard and a cloudformation validator into a new guard plugin.

The Go continuous delivery platform is going to be released as open source. It’s available free now, with the source code available shortly. Go is similar to popular continuous integration tools but with a strong focus on building pipelines. Will be interesting to see if this spurs wider adoption.

Working out ways of writing more testable code is a big part of adopting an infrastructure as code approach in my view. This post is a good example, demonstrating how to test custom facts in puppet, but the approach taken is general to anything you want to test.

A new Google+ community for people to talk about auditing in the context of organisations adopting devops practices. If you’re working in a regulated environment this might be of interest.

Sorry about the ads on this next post but it’s an interesting read. It boldly proclaims devops as being the biggest change in IT since outsourcing. It’s interesting to read opinion pieces aimed more at enterprise CIOs and viewing devops through the perspective of organisational change.

Another similar example, with a short case study of GE moving in parts away from specialists towards more full stack engineers. For people in very large organisations cases studies like this should help with convincing the boss.


BrightBytes is a fast-growing, mission-driven, start-up in San Francisco. We are building a data analytics platform to improve learning for millions of students around the world. We have an opportunity to build an awesome infrastructure and team.

IPC Media are looking for an Infrastructure Developer to help evolve our web hosting platform to the next level. If automated deployments and infrastructure as code is your thing, then we can offer you a good working environment and space to innovate. Email [email protected] or talk to @ells on ##infra-talk

Red Gate is looking for a Senior Systems Administrator to join their IS team to help deliver top quality IT services to Red Gate users. We’re looking for someone who is technically capable of looking after a predominantly Microsoft-based environment, and is keen to develop skills across a range of technologies. You should have a good understanding of cloud services and you’ve probably built a service or two on Amazon/Azure to prove it!


The second edition of dotScale, a one-day conference on Cloud, Big Data & DevOps will take place on the 19th of May in Paris. Early Bird Tickets and sponsorship opportunities all available now. And for the lucky devopsweekly readers here’s a 15% discount code off the ticket price: ‘DEVOPSWEEKLY’

Devopsdays is coming to Amsterdam on the 19th, 20th and 21st of June this year. The call for proposals is open now and closes on April 21st. I’m sure they will be looking for sponsors too.


Chef Metal is adding machines as resources to Chef, currently supporting Vagrant and EC2. We’ve got lots of good host level infrastructure as code tools, but very few cluster level tools that work as nicely.

Metrilyx is a web based dashboard engine for OpenTSDB, a time series database used to store large amounts of data. It allows for analyzing, cross cutting and viewing of time series data in a simple manner. The screenshots in the README give you a good idea.

Saboteur is a network fault injection tool that aims to simplify resilience and stability testing. Its core component is an agent that accepts commands over HTTP and configures its host’s network stack for various common fault scenarios including delays, packet loss and partitions.