Scaling a Kubernetes Platform across the Enterprise
by Jason O’Connell, Lead Engineer for Container Platforms at Macquarie Group
As an organisation begins using Kubernetes, real opportunities for transformation present themselves. Can developers’ daily lives become more frictionless? Can you increase the speed of delivering features to production? Can resiliency improve and tech-debt decrease?
Kubernetes alone manages running containers, but it is the opinionated developer experience that is built around Kubernetes that can truly unlock transformation. Builds, deployment pipelines, governance, security and many other concerns need to be addressed. And as the platform scales out and is used by more and more teams, the decisions on how users interact with the platform will have an impact, not just on developers’ day-to-day tasks, but on the organisation’s ability to deliver business value quickly.
At Macquarie, when we implemented the Openshift platform on Cloud we were determined not to just lift-and-shift, but to re-think how we delivered software with a focus on transformation. As teams migrated their applications onto the platform we knew we had the chance to clean things up, with two key objectives in mind:
- Making the developer experience as frictionless as possible
- Extracting common problems from every application, and solving them in a single place.
To meet these objectives we realised that the Platform team running Openshift had to do more than just run the platform, they had to work closely with developers to understand what was slowing them down; looking across teams to find where the same problems were being solved and extracting these common problems out.
This lead to the concept of the “Line of Innovation” which recognised that Product teams wanted to focus on delivering features and business value to production. Their innovative efforts were best serviced by being able to release fast, test and learn and iterate quickly. In the world of containers, they simply had to focus on everything they built into the container and not worry about the rest.
The Platform team was then not just a team to manage infrastructure, but a team of innovation in its own right. To build out solutions such as:
- Releasing multiple times per day into production
- Targeting releases to specific customer groups and channels
- A more centralised and holistic approach to controls
- Standardising approaches to monitoring, alerting and product support tooling
- Billing, resource management, vulnerability management.
The end result means there is now a dedicated Platform team looking at ways developers lives can be made easier, but more importantly from an organisational perspective, hundreds of applications can now be managed in a single way.
Standardisation and re-use brings us scale, with 365 applications onboarded over the past two years and 2000 pods now running in production. The capabilities built into the platform have accelerated our journey to the cloud and increased our ability to deliver business-value more quickly.
