DEVOPS WEEKLY ISSUE #107 - 20th January 2013
The start of the year has certainly got people busy. Lots of interesting tools and events in particular this week, so lots of opportunities to meet up with like minded folks wherever you are. Only two weeks to go too before the great Belgium devops roadtrip; PuppetCamp in Ghent, FOSDEM in Brussels and the monitoringlove hack day in Antwerp.
Sponsor
Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…
http://brightbox.com
News
Every wondered about what to log? This post, the 10 commandments of logging, should be sent to every developer or web operations professional and used whenever the question comes up. Well worth reading.
http://www.masterzen.fr/2013/01/13/the-10-commandments-of-logging/
Nice write-up of going along to the recent Devopsdays New York Event. Focuses on the community rather than the topics. I’m hoping the London event will have more of the big company and financial services folks mixed in with the startup crowd.
http://velohacker.com/2013/01/19/thoughts-on-devopsdays-nyc/
Ever deployed your application to a PaaS without realising you have failing tests? Here’s a simple pre-deployment hook approach using a new feature on Azure. The example is using node.js but the pattern should be applicable to anything.
http://codebetter.com/glennblock/2013/01/19/teach-your-website-new-deployment-tricks-with-azure-cli-1-of-xxx-running-mocha-tests/
If you like the weekly email newsletter format and have an interest in the Hadoop ecosystem them the just launched Hadoop Weekly might be of interest. Good luck to the author with keeping it going.
http://www.hadoopweekly.com/
If you use any sort of log aggregation or searching tools then this survey is looking for your input. 5 minutes of questions on size, volume, backends and usage patterns. I’ll look out for the results when they are published too.
https://docs.google.com/a/sematext.com/spreadsheet/viewform?formkey=dGNKUGVlSEw1RkdsWWZReExaVFhCeUE6MQ
Well put argument that configuration belongs to the environments and not the application. A good before and after developer experience report of going from using Powershell and a hand rolled configuration management setup to using Puppet on another project.
http://uttamkini.wordpress.com/2013/01/14/configuration-belongs-to-the-environment-and-not-to-the-application/
Jobs
Datadog is looking for a senior DevOps engineer to join the team. Datadog is a monitoring service that ingests and processes O(10^10) metrics every day. If you enjoy operations on a large-scale, real-time platform, if you live and breathe AWS, Chef, Python, Cassandra and the likes, do apply.
http://jobs.datadoghq.com/apply/qyGWBC/Senior-DevOps-Engineer.html
Events
QCon is back in London from the 4th to the 8th of March. Tutorials and lots of tracks, including several that should be of interest to readers of this newsletter. If you’re heading along you can use the code RUSH100 and save £100 off the price.
http://qconlondon.com/
10Gen have lots of Mongo related events coming up around the world and are on the lookout for speakers. Here’s a selection with more on the website:
MongoDB London - April 9 - CFPs due Feb 22
MongoSF - May 10 - CFPs due March 8
MongoNYC - June 20 - CFPs due April 19
http://www.10gen.com/events
Devops Con is being held in Herzilya, in Israel, on the 28th of January. A good looking range of talks no continuous deployment, devops culture, configuration management, logging and some interesting looking case studies.
http://www.devopscon.com/
Tools
Cubism.js is a javascript plugin for visualising time series data. It looks ideal if you’re building dashboards based on Graphite or Cube. The demo is very pretty and the focus on usefulness and usability is fantastic.
http://square.github.com/cubism/
Vaurien is a different take on the chaos monkey. It’s a transparent TCP proxy that can be installed wherever you’re making connections. The magic comes in with behaviours, which can be triggered with code and do things like introduce a delay or increase the latency of the connection.
http://vaurien.readthedocs.org/
Tatin is a light-weight in-memory HTTP state machine. It lets you store and retrieve volatile data using simple PUT and GET requests which might be useful for various orchestration tasks.
https://github.com/netmelody/tatin
Orc is described as a Model driven orchestration framework for continuous deployment. It combines a git based CMDB with mcollective to ensure applications are at the correct version and active on the load balancers.
https://github.com/youdevise/orc/
A (paid for) application for iPhone and soon Android, PilotSSH looks an interesting way of doing remote administration for common tasks. The GitHub repository of scripts is a neat idea.
http://pilotssh.com/
https://github.com/Geal/PilotSSH-scripts/