3 minute read

For those of you on holiday this week here’s a bumper set of blog posts, tools and presentations to keep you busy in-between the festivities. A big shout-out to anyone who is on-call over Christmas as well, here’s hoping nothing happens.


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


A great detailed post about logs. What they are, why they are important and how best to use them at scale. Lots in here, from distributed systems design, to data warehousing and ETL, to a good list of papers to read.

A presentation attempting to model devops, using lots of existing visualisations of what devops means and also looking at patterns that emerge from those models. Great stuff.

This weeks survey is the yearly Devops survey from PuppetLabs, Gene Kim and Jez Humble. 10 minutes of interesting questions (and a few prizes) and then a wait for the results.

Devops in large organisations always comes with interesting challenges. This post, with lots of concrete examples, talks through a devops journey inside a large bank. Focuses on the need for shared configuration management, and includes some mockups of automated change management tooling.

A 2 part series about using Docker as part of a Jenkins based build pipeline. The posts cover how and why to use containers in this way, and present some tools and code examples.

The slides and videos from the recent Flowcon are now available. Some great content on Lean, continuous delivery and in particular culture and how to change it.

Ignoring the silly name part way through this is a good post about designers and how they fit into a more integrated, collaborative team. Includes data from GitHub to back up the point which is interesting.

A good list of ten non-functional requirements for development teams, focusing on good operations. While I think this is good I think the topic is much broader, maybe top 100 next time?

I always go round in circles when choosing a new language to learn so it’s always good when people write-up their decision and thinking. Here’s a pretty solid argument why learning Java makes lots of sense for systems people.


We have built an operations and build/release tools team and are looking for fellow “devops” enthusiasts to join our ranks. Our team is the backbone of Product Development at Blackboard, enabling continuous integration/delivery, automated deployments and infrastructure as code. Our emphasis on monitoring and metrics helps shape our team’s actions as well as business decisions. We believe strongly in open source technology and contribute back as much as possible, especially as a result of our hackdays and hackathons.


The next Monitorama conference is going to be in Portland on May 5th to the 7th and a call for proposals is open now until the end of February. If you have tales of interesting failures and successes you’ve encountered in production then get submitting.


A very handy new Chef library, Chef Sugar provides a number of useful helpers all in one place to simplify your Chef recipes. So things like windows? and ubuntu_before_lucid? and installed_at_version? to name a few.

Vector is a predictive scaling tool for Amazon’s auto-scaling groups. What that means is it looks at historical data to guess whether you’re autoscaling group will trigger soon and triggers it early to avoid the rush. Read the full example in the README to get a clearer picture.

MIDAS is a Python framework for developing a Mac Intrusion Detection Analysis System. It provides a small library, along with tools for writing checks and doing local analysis. Early days but interesting to see more automation in this domain.