Automate all the things: A modern suite of tools for DevOps Nirvana
When doing things the right way is easy, you do the right thing often. When doing the right thing is hard you start to work around it. Doing things manually over and over is bad, scripting is better, having a set-it-and-forget-it service is best. This session is about how to make doing the right thing as easy as possible using modern tooling so you can stop working for your computers and start having them work for you.
This session will not be an in-depth review of any one tool but instead will provide a look at general best practices and how these small pieces can be fit together into a cohesive whole to close the gap between development and operations. The key is fostering transparency and making things exceedingly easy to use.
The tools outlined in this presentation will be described
Jenkins: Automated testing, code coverage, continuous integration, and reporting.
Sensu: Stop using Nagios; this is monitoring the way it was meant to be. Sensu was built for the cloud and not only monitors and alerts but also collects important metrics throughout your infrastructure.
Graphite: Graph everything and start making decisions based on data
RunDeck: Sometimes you need to run scripts or commands on your servers. With RunDeck you can use your central puppet or chef data to find the systems you want to work with and run simple commands or bigger orchestrated scripts.
Updatamator: Don’t wait until you log into a server to see there are updates and stop getting billions of emails from apticron or yum-updatesd. Updatamator will group your systems by the available packages, notify you each morning of the available updates, and allow you to upgrade them with a single click. Also, this is a shameless plug (I wrote it).
auth-proxy: All of these services mean having an awful lot of logins; that is without auth-proxy. Auth-proxy gives your team single sign-on access to all of your ops tools through o-auth from your google apps account.