4 minute read

I’ve spent the last two days helping to organise the excellent Devopsdays conference in London. From the feedback I’ve seen so far everyone had as good a time as me, thanks again to all the sponsors, speakers and attendees who made the event awesome. If you’re part of this community you should think about getting along to a near-by event, or even getting together with a few people to organise something. Putting lots of smart, passionate people together in one place really is a great way of learning.


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

Devopsdays London

Nice write-up of all the sessions, along with some thoughts about the format of the two days. Lots of links to slides, books and other resources mentioned in the talks and open space sessions.

One of the open space sessions I attended was all about community resources for learning about Devops. These notes from the person who ran the session are a great resource on their own, with links to events, podcasts, blogs, books and even this newsletter.

Slides from one of the Devopsdays London talks, this one from Niek Bartholomeus all about Devops for Dinosaurs. The topic is all about bringing devops culture and modern tooling to a large enterprise environment, based on a great in-depth case study.

Another of the Devopsdays London presentation that I’ve found, this one from Gene Kim all about how best to sell your boss or your organisation on adopting devops. Mapping business failings to developer failings looks a great way to get different people to understand the problem from their own point of view.

More notes from Devopsdays London, this time covering all the talks and ignites. The notes are drawn from tweets made during the talks, but you can follow along with the topics surprisingly well.

One of the ignite talks, entitled What if configuration management was created by game designers. Lots of interesting ideas, from rewards, high scores, virtual currency, cheats, health and more.


A few people might remember Flapjack, which aimed to be a next generation monitoring tool. Due to various reasons it has been superceded by tools like Sensu which have a similar focus. Well now Flapjack has been rebooted, aimed at being the part of the monitoring puzzle that deals with alerts, roll-ups and notifications. Well worth talking a look at.

Docker is the soon to be open sourced container runtime from the folks at dotcloud. It’s aim is to be a building block for building isolated runtime environments, using LXC under the hood to isolate everything and hopefully provide a general approach to specific tools like virtualenv, rvm gemsets, etc.

I’m finding quite a few presentations and videos about practical scaling tips for small companies, most of them quite rightly focusing as much on not doing some things and what to do. This talk from Ruby Australia focuses on Rails, but nearly all the advice should be applicable to any startup or small company.

Interesting presentation from the development team at SoundCloud about how they are migrating away from monolithic Rails applications to a running small services on the JVM as they grow. Lots of good tips and experience, in particular about making the cost of experimenting low - they are currently using JRuby, Scala and Clojure to find which fits best.

Some interesting survey results about barriers to adopting a continuous delivery model. Findings include organisational culture, lack of integrated development and operations, technical hurdles and a general lack of understanding. Worth reading some of the conclusions.


HotelTonight offers last-minute deals at top-rated hotels screened by the HT team and lets you book a room in seconds flat. In less than 2 years, we grew from 3 cities in the US to over 90 destinations in 10 countries. We are named one of the most disruptive services of the year by Apple in 2012 and named Top Developer on Google Play. We are looking for a talented DevOps engineer in San Francisco to build out a new stack on AWS using Chef and scale the system as we offer the service to more people in more places. Ideally, you can help us open source as much of this as possible.

Neo Technology, the people who make Neo4j, the world’s leading graph database, are looking for engineers to join the team. You are enquiring and thoughtful and have great Puppet and Ruby skills. If you roll with Python or Chef and have an open mind, we’ll take a look at your cookbooks and your indenting. You will help build and support Neo4j in the Cloud as part of a global team and work with leading-edge technologies: Puppet, AWS, Ruby, New Relic, and other tools from the DevOps community. You’ll get to help shape our global operation and discover the delights of Swedish Kaffebröd.


One of the tools mentioned in a talk at Devopsdays London was Deploykit, from Future publishing. The talk went into more detail but this was an attempt to make a generic deployment tool that uses message queues for much of the internal communication.

Vagabond is an alternative take on the Vagrant tool, but one based on LXC rather than virtual machines after the realisation that VMs can be quite slow. Maybe worth a look if you run Linux and find lots of performance or stability issues with virtual machines.