4 minute read

A packed issue this week, with lots of news, events, jobs and tools - something for everyone hopefully. I’m also noticing that I’m starting to include more critiques which I’m finding increasingly a good source of opinion. If you’ve been criticising something recently write a blog post and send me a link.


Devops Weekly is sponsored by Brightbox Cloud: the multi-zone cloud platform built for High Availability.

Try an SSD cloud server in 30 seconds with your £20 free credit…


DevOps: From Adoption to Performance – featuring Gene Kim

DevOps isn’t just for the “Unicorns.” Watch this on-demand replay to hear Gene Kim explain how you can replicate the DevOps practices and successes of the so-called “Unicorns.” Discover the lessons he’s learned from studying these companies and how you can apply them at your company!


A fantastic talk on the importance of operations, and how to grow an operations team, in a high-velocity environment. The video, slides and a full transcript are available.

Great notes from a recent talk about how Google built its datacenter network. Lots of details and observations about how the push to service based architectures fundamentally changed network requirements.

A good reminder to be wary of rushing to new techniques or tools without considering why previous approaches worked or what problems they solved. Looking specifically at potential pitfalls with microservices and container base systems.

A critique of the Dockerfile format used by docker to build images, including comments on layers, the registry and approaches to testing containers.

A good experience report of using Terraform with Azure, focusing on some of the differences between Azure and AWS and with lots of examples.

A good point on why individual log items are rarely enough when debugging a problem, and a discussion of a tool to help correlate items together.

An introduction to the rise of containers focused on the benefits to application developers. A decent run through of some of the advantages, new emerging tools and companies providing backing or support for containers.

Adopting Devops practices can be bottom up or top down or some combination of both, but without some buy in from management and without some understanding of the business goals in the development and operation team you can only go so far. Several good points on the theme of management and devops in this post.

If you’re running a web facing site or application you should be using a variety of HTTP headers aimed at making the web more secure. Based on this interesting report lots of people aren’t doing.

A good presentation about how to adopt a test driven development approach for infrastructure and operations work. Code examples as well as as discussion of strategies for implementing.

Another good presentation, this one on using CDNs as part of a security defence in depth. Lots of shoutouts to HAProxy, as wellas mention of a test suite for proxies/load balancers which is a pattern I’ve seen a few times recently. Good notes from the talk, with the video and slides at the bottom of the page.


We at trivago are looking for an experienced Site Reliability Engineer / Software Engineer to join the world’s largest hotel search and ensure a reliable, fault-tolerant production environment. If PHP is your mother tongue then we should talk!

Sainsbury’s Digital are looking for a DevOps Team Lead in London or Coventry. As well as managing a small team of engineers you’ll be in the thick of designing and building interesting distributed systems that scale to meet growing demand. Experience with common monitoring, configuration management and operating system stacks is essential as is a desire to approach DevOps practices in new and creative ways.

Flipboard is looking for a site reliability engineer leader to help us manage, maintain and monitor all aspects of the Flipboard production environment. We work exclusively in AWS, ensuring a highly available and performant Flipboard experience for millions of users daily. If you have a passion for working with large scale distributed systems and the challenges that come with them, come check us out. You will be leading the team that scales with the growth of a highly trafficked mobile and web app with concurrency in the millions coupled with services that consume the world of content and social signals.


Infra Bootcamp is an interesting new event in Houten in the Netherlands on the afternoon of September the 8th. A selection of talks focusing on automating Windows environments with various open source tools including Puppet, Chef and Openstack

Devops Sprint is a one day event for experimenting and hacking on infrastructure tools in Halle in Germany on the 30th of September. The organisers are collecting ideas for projects people would like to work on before the event.


dlayer is a handy tool for anyone building several docker images. It produces statistics about the use of layers, including how many layers are reused and how much space is being take up.

Battleschool is a development environment provisioning tool based on Ansible, similar to Boxen or Kitchenplan.

Testing your systems with inputs designed to cause problems is a good idea, but requires a good list of bad inputs. This list of naughty strings should come in handy.

Devops Weekly is sponsored by Brightbox Cloud: the multi-zone cloud platform built for High Availability.

Try an SSD cloud server in 30 seconds with your £20 free credit…