What is Platform Ops?

If you are looking to scale DevOps, Platform Ops is an approach you need to be aware of and may want to consider. It is picking up speed with organizations that realize that sporadic DevOps experiments, even successful ones, are not going to scale on their own. “How to Scale DevOps by Building Platform Teams” by Gartner is a good research note about this topic, which they are also going to cover in their 2020 IT Infrastructure, Operations & Cloud Strategies Conference. 

I loved how Bryan Finster, a Value Stream Architect and Walmart DevOps Dojo, explained the challenge very simply on LinkedIn: 

“ Teams choosing random delivery platform tools SOUNDS awesome, except when you are trying to make those pipelines safe, secure, auditable, and easy to train. Development teams should have good choices, but not choices for the sake of choice. Speaking as a long-time developer who now works in the delivery platform area for a Fortune 1 company, a centralized platform is a major advantage.”

Platform Ops is about supporting a culture change with the right organizational structure.

How do you do Platform Ops?

In order to adopt Platform Ops you need to create a dedicated team that maintains a self-service platform for the application development teams that includes everything needed to build an automated DevOps value stream, so developers don’t have to worry about infrastructure and operations tooling and can focus on developing.

What do I mean by DevOps value stream? Unlike CI/CD pipeline which only covers the process from introducing code changes to deploying to production, the value stream is a wider term for the process that starts from the business need and all the way to the customer getting value from your product. 

What are the benefits of Platform Ops?

Developers enjoy self-service, speed, and no need to be DevOps/infrastructure experts to achieve results fast and to automate their value streams all the way from planning to delivery to customers.

ITOps gets the ability to standardize and introduce guardrails through automation, for anything from cost to security and compliance, without being a bottleneck.

Is this ideal only in theory? The biggest problem is that it sounds awfully similar to a centralized IT approach, which is what made shadow IT come to the world in the first place.

Making it work requires balance, and every Platform Ops person that I talk to mentions this fine line.

How do you achieve balance in Platform Ops?

The development teams have to get enough freedom and flexibility to do their work, but on the other hand, the Platform Ops team needs to be opinionated enough to ensure that things don’t get out of control.

The bigger the organization, the harder it is to figure this balance out.

How do you make the platform work for developers?

In order to get collaboration from the development teams, you need to make the platform as flexible as possible when it comes to the choices that are important to developers, change their workflow, and affect their architecture decisions. Choice of continuous integration systems, a source control platform, where to keep artifacts, what type of artifacts, continuous integration systems, how to automate integration tests — if you try to be very opinionated on this front, it’s hard. 

The platform needs to be part of the developers workflow, so the developer experience has to be a main consideration for the Platform Ops team.

How do you make the platform work for ITOps?

Making developers happy isn’t the only concern. Platform Ops needs to be opinionated on things that really matter. Anything that has implications on cloud cost, security, and compliance has to come from the Platform Ops team, and there is less room for flexibility. Things like infrastructure automation, access control, secret management — these are the elements that result in unexpected cloud bills, in keys being exposed in public repositories, in data hygiene issues, in incorrect security configurations.

A Platform Ops team has to have high EQ. It’s a team that has to pick its fights, define its boundaries and coach other teams. But it is in a great position to be the bridge between developers and IT that many organizations are looking for.

I would be curious to know how you are using Platform Ops to scale DevOps, and your tips and tricks for a successful Platform Ops practice!

You can read more about Quali and how we help balance development speed and governance for DevOps here.