Charlie Murphy's True DevOps Stories
I've come to notice a pattern in DevOps tools and their adoption. First, they start by beckoning to you from across the room during a demo session at your local conference.
Inspired, you go ahead and install it at your organization. After all, everything looked so easy and revolutionary in the demo.
But wait - nothing but the most simple use case is working right! The documentation doesn't say anything about this error you're seeing in the logs (frankly, it doesn't say much at all). And what are all these files that suddenly appeared in /var/lib?
Still you see all of your friends (and the rest of the Internet) are praising it, so they can't all be wrong...right?
So you press on. You even feel a bit accomplished after 8 hours of Saturday patching heroics to fix a trashed node. It's time to write a blog post to demonstrate your expertise.
But the sensation doesn't last. Day after day, patch after patch, the failed RPM installations begin to wear on you. Regrettably, your once new infrastructure has made itself at home in your environment.
Months pass. You've reigned in most of the extra complexity from the distributed terror you've wrought upon the world, and have since moved on to the next shiny tool. Maintenance is something other people do, anyway.
Still, networks and hardware do fail from time to time and when they do, your creation reasserts its independence.
In conclusion:
