Before we start looking forward to Thanksgiving (I'm most excited for the stuffing), let's look back at some of the best DevOps posts from the last week. This week: release testing and risk management, customer feedback metrics, using Vagrant and Ansible, and a case study from Orbitz.
Steve Smith walks through the typical testing processes and different kinds of testing, and comes to the conclusion that release testing is a huge barrier to implementing Continuous Delivery. Release testing has very limited usefulness and of course, if it increases your time-to-market, it also increases the cost of production.
There's an awful lot of talk about decreasing lead times, becoming more agile, automating everything, and it's all about speed, speed, speed. One end goal of Continuous Delivery and startup culture that doesn't get enough press is the responsiveness to customer needs, though it's always implicitly stated. Since responding to customer needs quickly is so important, shouldn't we be able to measure our efforts?
MVB and Zone Leader Alan Hohn works at Lockheed Martin (no word yet on the official DZone private jet), and as part of a big company, he has to deal with web proxy server frequently. Here, he shares his experience working with these, and how to make VMs with Ansible and Vagrant work as smoothly as possible with them.
Case studies are always fun to read, because you see companies mentioned that you rarely think of being on top of their release processes, and it's always interesting to see how they came to their conclusions. Orbitz is another company joining the ranks of Netflix and CapitalOne. Their team has focused on adopting Enterprise Jenkins for CI and other tools for release automation and delivery. They may not be Amazon-level yet, but from a deploy every two weeks in 2013 to daily in 2015 is a huge leap.