Is DevOps dead?
DevOps is splitting up. Awwhhhh...Don't worry, the kids will be ok.
Developers are getting tired.?They no longer want to deal with operations.
The “you build, you run it” mentality is over.
Is it time to split development and operations again?
Developers are under a lot of pressure.??The software development lifecycle is getting more complex and more demanding.?Just like we do with everything else in tech, is it time to go back -- to a centralized operations team?
I mean DevOps is 15 years old….A lot has changed since then.
When it started, it was possible to build a super team of mythical creatures called full-stack developers.?
It is getting harder to find and retain these unicorns.
We are starting to see the push-back. We are starting to see tweets like this:
“Devs don’t want to deal with operational concerns, for the most part,”
“I am a dev and I don’t want to deal with operation concerns,”
“Devs and ops should work closely while having differentiated roles. The empathy between teams is the real point,”
It made sense at the time, but it is creating a serious bottleneck.
The challenge
While workloads increase, we are expecting developers to not only provide excellence with code, but also, monitor, secure and lead compliance.?On top of that we task them to build and maintain CI/CD pipelines.?And then we hold them responsible for all of this and measure them on getting code out quickly and safely.?Sure….
It’s too much.
The solution
We need to provide everything as a service to the developers so that we empower them with the right information and platform.
Let them build software, which is what they are good at.?They shouldn’t care unless something breaks.
But if we are doing the proper testing, we shouldn’t see these breaks in production, with their code.
As we provide them with everything as a service, developers can have full self-service access to development and test environments.?They can then take the infrastructure as code templates and provide them to the teams for production deployments where operations is responsible for production.
Kubernetes gets us there, but this still requires expertise to keep the infrastructure running and abstracting this away from developers.
The future
As we abstract the layers away from developers, we may see a resurgence of platform engineering and cloud engineers.?They will give developers the tools they need, with guardrails, to enable devs to do their best work.
A centralized operations team may emerge that removes friction for developers while giving them the dials to turn.?This centralized team will provide expertise, consulting and tooling support.
This centralized team will advise and consult with teams on best practices.
This team may eventually be called Site Reliability Engineering (SRE).?It may also be called Platform Engineering.?But at first it will be two separate teams.?They will be responsible for reliable platforms, most likely Kubernetes.?This will allow them to provide everything as a service to the developers.
I think we will see the following three dominant groups emerge over the next 2-3 years
Technology Manager at The Walt Disney Company
2 年Always enjoy reading what you have written Dale. Thank you