KubeCon + CloudNativeCon Europe 2019

Kubernetes runs the distributed replicas of an application completely independent with no mechanism provided for coordination between replicas. To minimize race conditions, and other side effects of distributability, applications are often deployed with a replica count limited one or use primitives like StatefulSet to influence pod scheduling stickiness. Unfortunately, these are not enough to create truly distributed applications that can gracefully react to changes in their environment at runtime. While the controller and operator patterns have helped, they tend to use raw API primitives that are not related to building distributed systems.

This presentation is a discussion that presents the features needed to create truly distributed applications that can react to changes in a distributed environment and including features such as leader election, synchronization, and coordination.

The popular sidecar pattern has influenced the way organizations integrate their infrastructure with Kubernetes. However, it can often be a challenge adapting the sidecar pattern for workload patterns like batch jobs that do not fall into the domain of Deployments and the traditional microservice paradigm.

This talk will describe the current problems and limitations with running Kubernetes Jobs in infrastructure setups that involve sidecar containers. We'll briefly cover how Spotify's use case with Jobs and sidecars conflicts with these issues. But more importantly, we'll cover what the roots of these Job and sidecar issues are, explain a few of the most popular workarounds in the community, and highlight the currently open Kubernetes Enhancement Proposal that presents a potential solution via new Kubernetes features.

The majority of production incidents for Pusher's platform team during 2018 had the same root cause... a broken config!

Kubernetes doesn't have a sense of versioning for configmaps and secrets and, in a lot of cases, updating configuration won't even be reflected into running pods. So how can you make sure that your pods are always running the latest configuration?

We came up with a solution for this problem in late 2018 which will be the focus of this session. What is Wave? What does it do? How does it work?