November 29 conference
Under the cut of the video of the conference reports - how Tinkoff.ru wrote their Infrastructure Provider BareMetal, how Kubernetes was raised in operation by Mail.ru Group, an interactive about the riddles of Helm with its RollingUpdate - and many other interesting things in the presentations of CarPrice, Eldorado.ru, Rosgosstrakh, Brain4Net, as well as a competition for future @Kubernetes speakers.
Thanks to the conference guests
Thanks to everyone who joined us - without you it would be a different @Kubernetes.
And here is the video:
Opening of the conference. Ilya Letunov, Head of Mail.ru Cloud Solutions platform
In the program, we sought to collect the widest possible range of options for using K8s in the Russian landscape. You are waiting for the stories of those who implement K8s in Production, create their own tools to work with it, collect errors and look for ways to solve problems, help others move to Kubernetes, develop as a K8s provider, and work to spread knowledge about the nuances of the technology. We hope that this will form for you in a single picture of the collective experience of the Russian Kubernetes community.
How Tinkoff.ru wrote their Infrastructure Provider BareMetal. Stanislav Khalup, Head of Attraction Infrastructure Group, Tinkoff
Stanislav Khalup has been the head of the attraction infrastructure group at Tinkoff.ru for two years now and often talks about Managed K8s and life in public clouds, but this time he will share what technological challenges he faced when building a Hyperscale Private platform from scratch.
How we transfer Mail.ru Group services to Kubernetes. Mikhail Petrov, technical manager of the Platform project, Mail.ru Group
Mikhail Petrov leads the Kubernetes Competence Center at Mail.ru Group and is responsible for the transition process to Kubernetes for all the group's services. In the report, Mikhail talks not only about our cluster and our pipeline, but also the skills that a person who starts Kubernetes should have, as well as about the transformation of processes in teams and the search for compromises.
Interactive βHelm through the eyes of developers. Rolling Update Puzzle. Dmitry Sugrobov, developer, Leroy Merlin
Kubernetes is the de facto standard and Helm is another default choice. But here's a simple question: what needs to be done to update the version of an application launched with Helm? Call helm upgrade? In fact, it's worth going a little beyond the tutorial - and you will immediately encounter a lot of nuances. In interactive mode, let's gallop through what Helm is in general, look at its architecture and storage device - and finally answer the question of what needs to be done to make the desired RollingUpdate still work.
Our evolution as a Kubernetes provider. Dmitry Lazarenko, Product Director, Mail.ru Cloud Solutions
Everyone is used to Mail.ru Cloud Solutions talking about how some complex things work in Kubernetes. In reality, we have been using Kubernetes for more than two years and have been offering it for about a year and a half.
Kubernetes in the spirit of piracy. Yuri Builov, Head of Development Department, CarPrice
Today, launching new projects on Golang and React, CarPrice confidently holds the helm of Kubernetes and remembers with a smile how they tried not to drown in the pursuit of βthe whaleβ on fishing boats with an elephant on board. This story is "not about driving a big beautiful ship on sails, but about a fleet of small, unsightly fishing boats - rusty in places, but fast, and so nimble and dangerous."
How we rolled K8s in Production eldorado.ru. Konstantin Rekunov, head of the Eldorado.ru IM operation group; Denis Gurov, Lead DevOps Engineer, AGIMA
Eldorado.ru use Kubernetes on their own cloud/hardware in a relatively small cluster and with a small number of services, but have already managed to get a lot of useful bumps. Colleagues share their experience of implementing K8s in Production, including touching on the topic of moving a cluster between DCs without downtime, which is interesting for many, and how it turned out to build a process for promptly diagnosing problems. The speakers paid special attention to practical cases of solving problems with the network, since these problems turned out to be very painful and non-trivial for them, and also tell what advantages they got from the implementation and what development prospects they see.
Implementation of OpenShift in Rosgosstrakh: from DevOps to Production Operation. Alexander Krylov, Head of DevOps Service, Rosgosstrakh
Being in front of a strategic fork - Kubernetes or OpenShift, the operation of Rosgosstrakh chose the latter because of Time to Market. As a result, we managed to build CI / CD using OpenShift, Bamboo and Artifactory and ensure the operation of services based on OpenShift in Production, as well as integrate them into the existing ecosystem of Enterprise tools used. From the speech, you will learn about the thorny path of integrating different boxed solutions that colleagues have gone through in recent years, and about the possible difficulties on the way for companies that want to repeat it.
Secure Kubernetes networks with eBPF and Cilium. How to network deeply at the kernel level? Alexander Kostrikov, DevOps engineer, Brain4Net
As an SDN company, Brain4Net strives for programmability not only in network hardware, but in the entire network stack. Kubernetes can also be programmed with more flexibility than just iptables rules. This allows you to make a tool like Cilium. Let's take a look at how Cilium uses eBPF to work with kernel and userspace code, as well as for networking, security, monitoring, and load balancing.
Contest "Ours at KubeCon"
As part of our @Kubernetes Ambassador support program, those who
Stay tuned
Follow @Kubernetes announcements, as well as other Mail.ru Cloud Solutions events in our Telegram channel:
Source: habr.com