4am on-call alerts that wake you only to restart an unresponsive web server. Developers, who can’t decide if five machines is enough to run their new application. Sudden traffic spikes that force you to spin extra service instances on Friday evening. Misplaced deployments. Frequent requests for yet another 50gb of network storage.
Does it sound too familiar? Are you wondering whether some cluster management system could take care of all of this boring stuff and let you focus on really cool problems? Then this talk is for you!
During this presentation I will describe Kubernetes features that help to reduce the maintenance cost of running containerised applications. After a really short refresher of basic Kubernetes concepts I will explain scheduling hints, probes, auto-scaling and automatic storage provisioning, and show how to use these mechanisms to make your deployments more robust and quickly adjust them to changing traffic or operational needs.
The keynote talk for this conference is about automation and this talk will show how to bring automation to production in the (probably) most popular container orchestration system.
view full program