DEVOPS WEEKLY ISSUE #108 - 27th January 2013
I’m really looking forward to the Belgium roadtrip this week. Puppetcamp, FOSDEM and the monitoringlove hack day one after the other. If anyone is going to be around Ghent, Brussels or Antwerp and wants to meet up let me know. I’ll try and get next weeks issue out when I can, but it will rely on me finding an internet connection so might be a little delayed.
Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…
Good in-depth walkthrough of building a Jenkins pipeline to build vagrant boxes with veewee and upload them to S3. Covers all the pitfalls and provides simple instructions. A useful guide if you’re doing something similar.
Inspired by a recent devopsdays event someone put together a reading list of devops related books. Only one overly technical book too, with lots about process and organisation change.
Lots of good talk about the role of being on-call recently. This blog post describes how a small startup have evolved their on-call rota and how they are now involving operations, developers and soon designers in the response.
Good talk from the recent devopsdays conference by Dave Zwieback all about hiring people into roles in a devops minded organisation.
Interesting post about the difference between composable and contextual ways of extending tools. Some good points for anyone writing tools with APIs or wanting their tools to be extended, and for anyone who hates Maven and wonders why.
Stripe is building next-generation infrastructure for online payments. We’re hiring engineers with operations experience to take our systems to the next level of scalability and reliability. If you love building high throughput, chaos monkey-proof systems; if you’ve built and maintained large software projects before; and if taking updates is a cathartic experience, then you should do-release-upgrade to Stripe. To apply, send an email to [email protected] with a description of your experience, relevant online handles, and anything else that would be useful for us to know.
Ledbetter is a relatively simple nagios to graphite integration. It takes the state of various checks and servicegroups in nagios and exports that data to graphite with a script you can run on a cron.
I like open source projects that make it clear what license they are released under. But copying a file around into your repo is easy to forget to do. Lice solves that problem nicely with a command line tool that knows about lots of popular licenses.
Ybot is a hubot inspired chat bot which aims to be a platform for exposing data and commands in a central manner. It’s written in Erlang, but you can extend it with scripts in various languages including ruby, python, shell and perl.
Ironfan is a Chef orchestration layer, providing lots of tools and knife extensions for managing a fleet of machines. What caught my eye this week was the walkthrough tool, which allows for using the chef DSL to define a set of virtual machines.
Shameless promotion of something I’ve been hacking on, very similar to the above ironfan example. A puppet module allowing for defining a set of virtual machines and load balancers from different cloud providers in the puppet DSL, and using an idempotent puppet run to ensure that they exist.