2 minute read

A bit of a mix of new tools, old tools, tutorials, presentations and introductions this week. Oh, and a new Linux based operating system with some ideas I’d love to see adopted elsewhere.


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


Graphite is a pretty powerful tool, but it’s not the easiest thing to get a good understanding of. This post covers tips for naming metrics as well as code examples of now to instrument a few Ruby classes. Useful whatever language you’re using.

All the sessions from devopsdays Santa Clara are already up, with full sessions and lightning talk videos available for your viewing pleasure.

Using libraries from sites like ruby gems or npm is an interesting security challenge. This post explains why, with a great social engineering example and also shows how you can run your own gem server.

Interesting analysis of the AWS EC2 API, focusing on the eventually consistent nature of some calls and the expected response times. Some good tips for writing or using EC2 clients, but also a nicely documented approach you could apply to other similar APIs.

NixOS is a new flavour of Linux, but what caught my eye was the package management system Nix. It’s based on functional programming which means that it can ensure that an upgrade to one package cannot break others, that you can always roll back to previous versions and that multiple versions of a package can coexist on the same system. It also supports multiple users and a built-in configuration management tool.

Systems are increasingly being built out of small services composed together. This post is a good introduction to why this approach has emerged and what some of the advantages and disadvantages are.

Nice presentation on the changing role of Quality Assurance and Testers within organisations taking a Devops approach. Lots of before and after examples and an attempt to explain why things should change.


Packer is a standardised tool chain for creating base images for different platforms, including virtualbox, aws, vsphere, etc. It’s ideal for automating this sometimes tricky task. The 0.1 release doesn’t support ever hypervisor or provider yet but the plugin architecture means it can do in the future.

Ever wanted to use a Chef cookbook but not wanted to alter maybe one line to make it work for your setup? Chef rewind will modify some properties of a resource, during the complile phase, before chef-client actually starts the run phase.

Flapjack is a distributed monitoring notification system I’ve mentioned previously. Getting up and running with all the bits however could prove tricky so here’s a way of building omnibus packages for it. It’s also a good example of using the omnibus tool.

If you’re looking for a simple provisioning tool in Ruby then Orca might be of interest. It’s aimed at a space between deployment tools like Capistrano and configuration management tools like Chef and Puppet, and provides a small DSL which deals with simple cases.