Hope everyone had a great holiday season, and welcome to the first DevOps weekly of the new year! This week, we're looking at infrastructure patterns, automation goals, why refactoring shouldn't be a major project, and how poor DevOps can harm your business.
We mostly cater to software developers here on DZone, but it's important to look at the infrastructure side of DevOps as well. David has put together a really in-depth article on four patterns: pipelines, hashed cache, retries and fallbacks, and pipeline in a loop, complete with some code examples.
Automation is often seen as necessary because it saves time that developers could be spending on other tasks, rather than pushing buttons. However, as John Cook (and xkcd) points out, sometimes automation efforts get out of control, and the prospect of "paying for itself" in time looks more difficult than planned. But even if that's the case, automation can save mental energy to give developers a break, or keep them in a zone longer, and there's something to be said for that.
How many of us have started to accept "refactoring" as a term for "rebuild everything"? MVB Eric Dietrich presents a sanity check on what refactoring actually entails, and why refactoring should focus on very little changes that shouldn't require any effort at all, like taking your shoes off before you enter a house.
We hear about massive IT failures on the news all the time, and this case study from DZone user Ron Giron looks at the way Oyster's ticketing system for the London Underground failed, keeping barriers open and losing thousands of pounds of revenue, and what can be done to prevent similar failures. In this case, DevOps was implemented poorly, and unfortunately, even reverting to the "old way" of doing things in just one spot can spell disaster.