Abstract:Before deploying Kubernetes to production, there are complex choices to make upfront: which cloud platform, which Linux distribution, what are the best tools or methods to manage and monitor your app, and more. We’ll kick off this session by going through some of these decisions, and highlight a few options to consider for each.
From there, we’ll look at a real world use case, where a small team went through the process of deploying a production service on AWS (full disclosure: it was one of our own internal product teams). At the time, we had limited exposure to the Kubernetes deployment process. We’ll give you a “behind the scenes” look into our team’s journey.
Our initial use case was simple: build a few microservices backed by Amazon Relational Database Services (RDS), and run them on Kubernetes. Our goal was to have the ability to run the cluster in any environment, from Amazon, Google and Microsoft public clouds; developer laptops; and dedicated on-premise deployments.
Join this session to learn how to make Kubernetes deployment easier - plus a few fun (potentially embarrassing!) stories as we made our way through the Kubernetes production experience.
Outline:
- Organizational challenges for adoption of Kubernetes
- Refresh on Kubernetes cluster architecture
- configuration challenges
- infrastructure requirements
- Decisions to be made when deploying a production cluster
- monitoring
- networking
- storage
- logging
- Variety of existing community projects
- What we ended-up doing
