When Steve arrived at Mettle he was greeted by manually provisioned clusters with nodes that were multiple years old and people were scared to change them. This talk will show you the journey Mettle is on to go from this setup to how they now see clusters as being totally ephemeral. Steve will talk about the lessons learnt along the way, the tools being leveraged and what’s left to do to complete this journey.
With Kubernetes, autoscaling is made easy, but there are still traps waiting for incautious. How to ensure high availability in shared large scale cluster – a journey through scheduling, load balancing and meaning of HA in Kubernetes.
GitOps got popular in the past few years, as kubernetes became the de facto standard for containers orchestration. Applying gitops to kubernetes is rather “easy”, since all resources are managed by kubernetes through yaml files. However, in enterprise environments adopting gitops is more problematic, since we usually have a melting pot of IT systems: cloud services, on premise cloud native services, on premise legacy services and systems etc. etc.
Kubernetes enables a possibility to develop cloud native microservices or decompose traditional applications making them more technologically advanced with the help of containers. But there is always room for improving 3 Ps – Performance, Price and Provisioning. Currently, most of the Kubernetes solutions are offered on top of VMs that brings a set of limitations while scaling and extra complexity while maintaining the workloads.
Never miss our news!
We love email, and you should too. Your inbox is where you'll find all the latest announcements and deals.
BONUS: Subscribe & get previous conference video archive!