DEVOPS WEEKLY ISSUE #128 - 16th June 2013
Velocity kicks off in just a few days. For all those heading to Santa Clara first off have a great time. Second, please email or tweet @garethr with any stories from the best presentations and links to interesting new tools.
Sponsor
Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…
http://brightbox.com
News
Great presentation talking about the relationship between developers and operations in terms of trust. Good examples and a few tips on how to build trust in people and organisations.
https://speakerdeck.com/sameaton/trust-in-devops
Ansible 1.2 has been released, adding a concept of roles to the configuration management tool as well as more modules and better variable support. The new documentation sites also looks much better and should make getting started easier.
http://blog.ansibleworks.com/2013/06/10/ansible-1-2-released-and-ansible-1-3-begins/
http://ansibleworks.com/docs/
Ever wanted to know when new versions of artifacts are released on Maven Central? If so then Artifact Listener is for you. Tell it which artifacts you’re interested in and it will email you when new versions are shipped.
https://www.artifact-listener.org/
Versioning software is one of those things that helps hugely when done well or can be completely meaningless or even misleading if done badly. So this document on Chef cookbook versioning is both a good idea and nicely done.
http://chef-community.github.io/cvp/
Interesting article on the pit falls in convergent configuration management systems, picking on Puppet’s exported resource approach and showcasing an alternative solution.
https://devblog.timgroup.com/2013/06/14/exported-resources-considered-harmful/
It’s easy to build applications and systems that will be difficult to scale when the times comes. Here’s a simple list of 10 things you should avoid that will cause problems later.
http://highscalability.com/blog/2013/6/10/the-10-deadly-sins-against-scalability.html
Jobs
Atlassian is looking for a senior Hosted Operations Engineer. If you dream about Linux, can write Python or Ruby in your sleep, and have an obsessive-compulsive attention to reliability, we’ve been looking for you! If you want to work in a casual environment with like minds, let us know who you are.
http://www.atlassian.com/company/careers/jobs/listing?org=ATLASSIAN&cws=1&rid=655
Enstratius is hiring sysadmins and developers with experience in distributed systems, cloud computing, and/or infrastructure automation. We have frontend (Javascript, Groovy), backend (Java, Python), customer/installation facing (Python, Chef, MySQL, Riak), and QA positions. Remote is a possibility. If you are looking for exciting work in a fast growing field, send an email with resume to: [email protected]
Tools
One area of monitoring that is occasionally discussed but that few people have implemented is anomaly detection. The nice folk at Etsy have just released Kale which might change all that. It can consume data from Graphite and provides both anomaly detection (skyline) and can discover correlations between an anomaly and other metrics (oculus).
http://codeascraft.com/2013/06/11/introducing-kale/
https://github.com/etsy/skyline
https://github.com/etsy/oculus
More vagrant providers keep emerging, this one allows you to manage SmartOS zones, including memory and disk limits and the networking configuration.
https://github.com/joshado/vagrant-smartos
Even more vagrant tools; Vagrant Multiprovider Snap provides a single command line interface to taking snapshots and rolling back to a previous state. It currently supports both Virtualbox and VMWare Fusion providers.
https://github.com/scalefactory/vagrant-multiprovider-snap
Log2Viz is an interesting looking project from Heroku which gives you some simple tools to treat log files as data streams. Nice simple implementation of using the Sinatra streaming API and server sent events.
https://blog.heroku.com/archives/2013/3/19/log2viz
https://github.com/heroku/log2viz