How DevOps Adoption Is Changing

To gather insights on the state of DevOps, we spoke with 22 executives at 19 companies implementing DevOps for themselves and helping clients to implement a DevOps methodology. We asked, "How has DevOps changed since you began using the methodology?" Here's what they told us:

Adoption

• As I talk to customers and prospects there’s greater awareness of DevOps and what it can do. It’s being taken more seriously. There is sufficient proof of organizations doing well with DevOps. This is a business process change.
• A survey we just conducted reflected the insurgency within companies to the way things are done. One year ago, it was, “what is DevOps?” Today there’s a common understanding with a desire to know how to scale.
• CD has become mainstream. Microservices are more commonplace and are a good way to be successful with DevOps. Being in the cloud gives you more flexibility. Containers are becoming more mainstream. Function as a service is a helpful way to solve scaling issues.
• People are beginning to understand the benefits of DevOps. Best practices have been solidified. Allows you to get code from developers to customers in a fast and secure way.
• 1) Beginning to just do with new projects. Today DevOps is a hard requirement. We’re in an operational race. 2) More focus on time to value – run a cycle, ship a single iteration. Ability to work across the entire pipeline. 3) DevOps engineer is a very common role today. Systems administrators and developers are converging. It’s blurry across the entire pipeline. Get visibility with tools to make everyone’s job easier.
• Three years ago, the State of DevOps Report showed the market cap growth for DevOps companies outperformed the S&P 500. Now finance is asking why we can’t release every 11.7 seconds like Amazon. There are new tools and apps for the customer. See HBR on the haves and have-nots of IT.
• The velocity of DevOps uptake over the last 12+ months. Amazon giving in to Kubernetes was a huge sign.
• Adoption of DevOps is becoming more prevalent in the more traditional enterprise space. We see it in our financial services customers. 
• DevOps has meant many different things over the past five years. At first, it was about the speed at which you can develop. Now it’s a soup-to-nuts pipeline. There is a greater definition of goals and KPIs – release chain velocity, ability to ingest change, mean-time to resolution. DevOps has been elevated to a level where management is watching its performance.  
• The movement to the cloud has forced the adoption of DevOps because development and operations get blurred in the cloud. The wall between the two is thinner.

Testing

• 1) Continuous testing. Testing becomes a center of excellence. To bring code into production in minutes, you need to infuse testing into every step of the SDLC. Performance testing is ongoing so you’re not waiting for load testing at the end. Acquired Blazemeter for cloud-based performance testing. It’s based on open source. We’re able to change the paradigm around performance testing as part of JUnit testing. We’re able to start small and scale. 2) Security testing in the SDLC so security is part of all other tests. 
• 1) The technology as there are many subcategories with solutions providing a breadth and depth of automating the process into production. 2) Testing has exploded in value and availability. Identify your complete testing suite before building. 3) Still cultural – how to get teams from development and operations on the same team, with the same metrics for success, with the incentive to collaborate. Best in class cultures collaborate, communicate, and are outcome focused.

Speed

• So many tools are hitting the market every day. It’s impossible for a customer to keep track of what’s out there and what the best solution. Organizations are willing to let teams be more creative and they see the benefit of doing so. 
• The pace of change with some companies releasing 50 times a day.
• Automation allows you to deploy regularly and quickly. Work with the QA team for push-button deployment. It takes two minutes and is scalable. Identify the issue, fix, deploy to QA to improve the efficiency of the team doing QA so they can focus on their job versus managing infrastructure. Consistent improvement. The window of error is always closing. 
• As we rapidly grew our business, our engineering team grew from a couple of engineers who did everything to dozens of engineers organized into several teams building and running our production systems. Initially, the challenge was keeping our brand-new systems running with a small number of customers, but in a relatively short space of time it became apparent that we needed to manage a rapidly growing server fleet (now over 1300 servers) without significantly growing the team. Moreover, customer demand has led to a need for a very high pace of change, which led us to move from scheduled releases to continuous delivery. 
• Since we’ve implemented the DevOps methodology, we’ve been able to release quality updates at a faster pace. This is due to the streamlining of internal processes associated with development, testing, and deployment of new features, enhancements, and bug fixes.

Other

• Adoption is different in every organization. Companies are still experimenting on what will work best for them. There are no best practices or standardization since DevOps implementation is different for every company.
• More vendors are combining features putting more DevOps services under one roof. We’re still early in the process and it is still evolving.