In a recent blog post about the importance of a DevOps Platform, GitLab CEO Sid Sijbrandij outlined four phases through which organizations frequently travel as their practice matures. It’s a painful journey we see again and again when we meet new customers. It spans every industry and every company size, and it’s the most mature DevOps teams with the most at stake who’ve felt the most pain.
Historically, if you wanted DevOps to work, you had to be prepared to pay for it. Just managing the backbone of DevOps – the toolchain -– has been a grind. Your “Jenkins Team,” your “GitHub Team,” or even, as one of our customers described, your “Duct Tape Team” (designed to hold it all together and patch holes), added no end value beyond keeping everything from falling apart. That’s a lot of investment to keep the lights on.
It’s a hard commitment to swallow, and the truth of it is that you shouldn’t have had to. A big part of the problems behind many “low-performing DevOps teams” stems from a poor set of tools for the job. Broadly put, on behalf of the DevOps tool industry: It’s not you, it’s us. The industry created many of these problems because we were thinking small and building to match.
As a philosophy, DevOps is pretty new, and it’s evolved very quickly. That rapid evolution has meant tremendous transformational opportunity, but building for the present left many tools, and the processes behind them, obsolete as soon as they hit the market.
DevOps toolmakers have long been focused on solving discrete, easily understood problems (“BYO DevOps” in Sid’s blog), while DevOps has always aimed at solving bigger problems and looked to a more collaborative, productive transformation. You knew that when you tried to calm the chaos by implementing standards (BIC DevOps). You knew that when you tried to Frankenstack those tools into a servant of your larger ambitions with DIY DevOps integrations. But in the end, tools were creating almost as much work as they automated.
It makes sense when requirements are evolving so quickly. In 2011, when GitLab offered just a repository and issues, we couldn’t have foreseen Design Management or ML Ops, but ten years later, they’re a key components of a movement toward a DevOps Platform for everyone. And that’s the point of the DevOps Platform Era (Phase 4). We’ve iterated our way to a place where we can replace blockers with enablement, and *support* your efforts instead of increasing your burden.
[Stop paying the “DevOps tax” by moving to a DevOps Platform. Here’s how]
This isn’t unexpected. Every technology reaches this inflection point as it evolves. In the not-too-distant-past, customer relationship management (CRM) required a portfolio of sales force automation and marketing automation tools, commerce engines, app servers, analytics engines, and huge amounts of data integration to make it work. Now we have SaaS-based CRM solutions with a single monthly fee.
While GitLab has always focused on delivering a DevOps Platform as a single application, we're excited to see the industry as a whole shift to a platform mindset. Late last year, Gartner released its vision of the DevOps Value Stream Delivery Platform in a new Market Guide, in which we’re happy to be a representative vendor, and we’re excited to watch their coverage grow.
We’re also excited to hear how a DevOps Platform benefits our customers in concrete ways. In our 2021 DevSecOps Survey, respondents told us a DevOps Platform resulted in better DevOps, improved collaboration, easier automation and expanded visibility and traceability. Or, as one survey taker said, a DevOps Platform “gives us reduced mean time to recovery (MTTR), quicker time to market, reduced lead time for fixes, and fewer change failures.”
DevOps hasn’t stopped evolving, and neither have we, but we’ve reached the point where we know how the pieces need to work together, and we’ve built a platform to support it. To see for yourself, try GitLab Ultimate for free!