The power of community, and doing things together as that community, was the overarching theme in much of KubeCon/CloudnativeCon presentations.
We were incredibly lucky to be able to attend, experience and learn from so many Cloud Native experts and developers over the past week.
From the keynote by Tim Hockin of Google and Kal Henidak of Microsoft Azure on the importance of "project over company" to Ian Coldwater's exhortation for developers to talk to people and learn from others' perspective, the thread of community power was present on the presentation stages as much as it was in the hallways, corridors and afterparties.
Kubernetes as a technology has enjoyed incredible success, and it's due to the open source nature and various companies all contributing resources to its development.
The number of speakers and attendees commenting that the size of KubeCon 2019 was larger than anything they could have envisioned even two years ago speaks to the excitement and growth within the community. We certainly could not be offering the world's first managed k3s platform without them - and you, our users.
We wanted to do a wrap-up of this exciting event, celebrating friends and collaborating partners, and looking into the future - and to KubeCon Amsterdam 2020!
The first thing we wanted to highlight was the announcement of the General Availability release of
k3s, the lightweight Kubernetes that underpins our managed Kubernetes service and allows our users to spin up clusters in under two minutes.
Our CTO Andy Jeffries wrote this post on
k8s when we launched our managed Kubernetes offering, and we have since kept pace with releases, with v1.0 being available within a day of the release announcement.
Exciting day for @RancherLabs as we celebrate general availability of <a href="https://twitter.com/hashtag/K3s?src=hash&refsrc=twsrc%5Etfw">#K3s -- the industry's lightest weight #Kubernetes distribution purpose built for the #Edge, now ready for production! Thank you Arm and @CivoCloud for partnering with us on the launch https://t.co/pnuoec8bjy— Rancher Labs (@RancherLabs) <a href="https://twitter.com/RancherLabs/status/1196797413828308992?ref_src=twsrc%5Etfw">November 19, 2019
Rancher Labs have shown incredibly exciting technology, from
k3s itself to
Rio, which we cannot wait to have a proper play with on our platform.
The interest in and popularity of
k3s has been astronomical as measured by GitHub stars, and there is more to come: Project Longhorn is expected to hit General Availability in the new year.
Bonus props to Rancher for including Civo in their presentation at KubeCon!
Databases as a service (DBaaS) are another exciting piece of the Cloud Native landscape. For a long time, the accepted thinking has been to keep databases out of clusters, but that thinking is changing.
We were especially impressed by the folks at Vitess demonstrating the abilities of their cloud database system, compatible with MySQL syntax, which underpins the likes of Slack and the Chinese e-commerce site jd.com. Vitess is now a graduated project of the Cloud Native Foundation, meaning it has undergone strict auditing and resiliency assessments.
While PlanetScale offers enterprise cloud native database solutions, their Vitess project is entirely open source, once again showing how the community comes together across company boundaries. The graduation of the Vitess project also means that it is not a siloed black box, but a community-enabled, community-audited product.
Kubernetes as an orchestration platform is intended to help with managing complex systems and allowing for self-healing, failovers and scaling to demand.
We enjoyed several very interesting presentations that looked at what happens if you kill parts of the service underpinning your app, both at a theoretical, testing level, and at the level of Chaos Engineering – the intentional disruption of service in production.
Learning from the experience of companies like Airbnb with their "Ten Ways to Blow Up Your Kubernetes" gave food for thought. Again, the value of wider experience was clear: You can build the best system that you can think of, but external factors or users may poke unexpected holes in it quickly!
Containers and continuous delivery go hand in hand. Being able to deploy a rolling update through your cluster means no downtime for users. Maintenance windows, where a service used to be unavailable at a predetermined time for all users, are a thing of the past.
Administrators can now deploy entire apps with a simple
git push thanks to GitOps. A really interesting take on the GitOps model was shown by the Argo CD Project from Intuit, which allows for declarative continuous delivery on Kubernetes through the use of a Git repository as the authoritative source of information.
If your interest is piqued, check out the guide Alejandro JNM wrote to deploying Argo CD here.
In terms of security, Ian Coldwater's keynote call for cooperation and awareness of mindsets other than our own – as attackers come in all manner of guises – highlighted the need for the wider Kubernetes community to be aware of security throughout the development process.
When you think like an attacker, you are more likely to bake in security mechanisms, meaning that your apps are less likely to be vulnerable or prone to failure because of user action.
Know what you're running in your clusters, and understand it well. Understand how your tech works, and think what an attacker would see when they look at it. #KubeCon19 - @IanColdwater Keynote pic.twitter.com/x3eLZ1wkAM— Civo Cloud (@CivoCloud) November 21, 2019
We were thrilled to be able to participate in KubeCon North America 2019, and are thoroughly looking forward to Amsterdam in 2020, where we are looking to have more of the team on site.
The Kubernetes ecosystem is an exciting place to be, and is made all the better by the community of people working in it, looking across job roles and company boundaries to create something more than the sum of its parts.