The DevOpsDays conference came to Zenoss’ and my home new hometown of Austin, TX over the last two days. Zenoss was one of the sponsors for the event, and I was fortunate enough to attend.
Going back through the blog, you can see that we’re no strangers to talking about DevOps. There are many reasons for this. As a software company focused on systems management, we epitomize DevOps at least in the literal semantic sense. More realistically, as we’ve built Zenoss to be the best system for monitoring a wide variety of resources from routers to firewalls to load balancers to storage to servers to applications, we’re very often pressed hard against the walls that exist between departments in large organizations.
Zenoss works best when shared. Due to the normal departmental firewalls and battles most of us have grown accustomed to over the years, we have our share of users who are using Zenoss just for monitoring routers, just servers, just VMware or something along those lines. That’s great by itself, and an even better start to something more, but the value only grows when all of the resources can be monitored together. As the meme goes, “monitor all of the things!”
Where this “best when shared” philosophy is true with Zenoss Core, it’s even more important for our customers with Service Impact. In this case, Zenoss can use a holistic view of all of your infrastructure to automatically discover dependencies and provide accurate root cause and service impact analysis when things go wrong, or just when things are in danger of going wrong.
So the DevOps philosophy of breaking down barriers, eliminating silos, getting everyone on the same page and singing Kumbaya, is music to our ears. We’ve been trying to crack that nut for years.
John Willis provided a great kick-off to the conference, and I feel like it set the tone very well for most of the talks during the two days. Back in 2010 he wrapped up what he felt DevOps was about in CAMS: Culture, Automation, Measurement, Sharing. To me, this conference was very focused on the culture and less on the others. Over time, this seems to be the general trend in the DevOps movement with automation and measurement fading a bit mostly due to what appears to be tools fatigue.
Pete Cheslock followed with a talk directly related to what turned out to be the biggest trend for the conference. Simply put, everyone is hiring. If you’re looking for good DevOps talent like we are at Zenoss, good luck! If you have good DevOps talent, watch out! I stopped tracking it at some point, but I think everyone who had a moment on stage rolled out the hiring red carpet.
Pete’s talk focused on how to retain the people you have, because as he so eloquently put it, “hiring people sucks.” I believe he was the first to reference Gene Kim’s book, The Phoenix Project. However, it was a very frequent topic and from all accounts should be mandatory reading for anyone in the field. Gene was even there to give a well received talk and sign the book.
Pete’s talk boiled down key concepts in a short time. Take money off the table as much as possible. Monetary adjustments are a a short-term fix at best. Understand what really drives people: autonomy, mastery and purpose. Open source projects, unscheduled time, and conferences and other forms of continuing learning being some example ways to satisfy these drivers. The over-arching strategy should simply be to treat people like the adults they are.
I could go on about some of the other quality talks by Dominica DeGrandis, Gene Kim, Jeff Hackert and others, but my best advice is to watch what’s coming out of the DevOps movement and these folks specifically. All of this could be just a blip on the radar, or the beginning of something big.