All Things KubeCon and CloudNativeCon Seattle 2018

It’s the second time the conference was hosted in Seattle. This time it was huge!

Highlights:

  • Kubernetes dominates
  • Kubernetes is hard and complex for application developers and there’s a need for a higher level abstraction
  • Knative is trying to provide a higher level serverless abstraction for Kubernetes
  • GitOps is gaining more traction
  • Service Mesh is everywhere (and Envoy is the king)
  • Service Mesh is not ‘fire and forget’ (be prepared to deal with its complexities and problems)
  • Unifying metrics, logging and distributed tracing (Grafana and OpenCensus with slightly different focus)
  • Grafana is adding support for logging (and tracing)
  • Debugging distributed container and Kubernetes based applications is getting easier (thanks to Telepresence, eBPF, VS Kubernetes extension/draft connect, Squash and other tools)
  • Distributed tracing is going mainstream, but it’s still too fragmented (Expedia made things even more confusing with Haystack :-))
  • Jaegar is adding new features to make distributed tracing for large scale systems more usable
  • NATS is maturing its cloud-native messaging capabilities
  • Rook is trying to address the cloud-native storage needs in Kubernetes
  • Prometheus scalability, high availability and long term storage is still tricky with multiple competing solutions (Cortex, Thanos, M3)
  • Kubernetes Operators and CRDs are hot (allow extending Kubernetes)
  • Kubernetes security is getting more visibility (especially after the recent Kubernetes vulnerability)
  • Kubernetes autoscaling is still maturing (cluster autoscaling and pod autoscaling with VPA/HPA)

Take a look at the schedule to see the slides and videos. There’s also a YouTube playlist. It will take a while to see everything though :-)

There are quite a few interesting post-conference blog posts. Read them first if you are not sure where you want to start exploring the conference talks.