The final day of KubeCon 22’ did not disappoint. The speakers shared amazing insights that will help our customers achieve more. I wanted to share my experiences at the conference to bring value to all the engineers that are utilizing Kubernetes in their day to day lives. For any questions or to speak to our team of Foghorn experts, contact us here.
In the keynote: Nurturing The Whole Project – Josh Berkus, Community Architect, Red Hat & Catherine Paganini, Head of Marketing & Community, Buoyant talk about the things you don’t think about initially that are very important for projects that are very important. You need to deliberately create a welcoming community, publicize your project, manage contributors’ expectations, motivate them to take on more responsibility, develop project policies and processes, and more. With so many things to tackle, it can feel overwhelming. Fortunately, lots of people have done this before and with this talk the CNCF and TAG Contributor Strategy are here to help.
In the second keynote: How Developers Help Scale Kubernetes Security – Connor Gorman, Senior Principal Software Engineer, Red Hat shares how properly securing applications deployed onto Kubernetes is a shared responsibility. Security teams define organizational policies that improve security posture while developers implement those policies through good security practices, keeping images up to date with the latest vulnerability fixes and configurations that follow the principle of least privilege. The declarative nature of Kubernetes allows security to be deeply integrated into development workflows. This integration empowers developers to be security stakeholders and scales the remediation of security issues. In this session, Connor explores concrete ways and best practices for integrating security into both CI and CD.
During the last solid keynote Push It to the Limit: From Canary Deployments to Canary Clusters Henrik Høegh, Platform Engineer at Lunar shared a method for making production clusters truly disposable, which took three years to achieve. They went over how today any engineer at Lunar can fail-over the entire production platform in 40 minutes, they did this with deep integration with the infrastructure provider as well as creating some custom operators. and moving most of the state out of the cluster the company is now in a position to make disaster recovery a day to day operation not just something tested once a month or once a quarter.
In the talk TikTok’s Story: How To Manage a Thousand Applications on Edge With Argo CD Bytedance shared a case study on how TikTok manages its global edge clusters with Kubernetes and operates continuous delivery with Argo CD, within this talk there was a lot of insight into the challenges TikTok faces with scalability with around ~100 edge clusters ~150k CPU’s and around 3000 different applications running in the clusters. with caching and traffic acceleration, they show people how to deploy and manage these kinds of services using argocd, with a nice demo to boot.
Laurent Bernaille & Elijah Andrews from datadog shared with the world a talk named Logs Told Us It Was DNS, It Felt Like DNS, It Had To Be DNS, It Wasn’t DNS in which they go about 4 weeks of pain and suffering trying to diagnose an issue which looked like it was DNS related. a pretty solid deep dive into the real problem and how it was resolved eventually by only removing 3 lines of code.
Moving on to Show Me Your Labels and I’ll Tell You Who You Are by Cisco showed one of the underestimated benefits of Kubernetes being the standardization of labels. Of course in every provisioned Kubernetes system, you have the ability to mark the assets with taints, tags, or labels. Prometheus raised the stakes and built a whole metric system on labels. The concept was so successful that more and more tooling tries to benefit from it. Modern logging, tracing, and metric systems have at least one common characteristic: they have labels. In this talk, Sandor demonstrates how to fully exploit labels with tools that are available already at your hands—how to correlate different inputs, transform logs to metrics, and more!
Thanks for joining me at KubeCon. If you haven’t already, here are links to the original day one & day two posts to fill you in: