In May 2019, the CNCF(Cloud Native Computing Foundation) held its annual KubeCon together with the Continuous Delivery Summit. This year the event took place in Barcelona and had approximately 8,000 attendees from all over the world. The event’s main focus was around customizing and extending Kubernetes, everything that has to do with containers and how they integrate and work with other DevOps open source tools. The event held more than 300 sessions, on different topics like AI/ML, container security, CI/CD, service mesh, networking, serverless and more.
For the first time, AllCloud attended both of these events. Yossi Jana, container technical lead and DevOps team leader, and Tal Helfgott, DevOps engineer and CKA certified, both brought back a mass of new knowledge they learned about Kubernetes.
I had a few questions for them. Let’s hear about their experience at the conference!
Q: First, Tell Us a Little Bit About Yourselves.
Together we currently have experience with more than 10 Kubernetes clusters running in a production environment. Within our organization, we are leaders in this area and a part of the Containers/Kubernetes ‘Domain of Expertise’ team. In addition, we have led several Amazon EKS workshops at FLOOR28 (AWS office in Tel Aviv).
Q: Why Did You Choose to Attend KubeCon?
When talking about Kubernetes or any open source platform in general, there is usually more than one solution for any given scenario. In fact, the CNCF is running more than 650 projects, divided into several different categories, and each category has multiple solutions to choose from.
Imagine that your goal is to build a perfect puzzle from just 10 pieces but there are more than 1,000 puzzle pieces on the floor. How do you find the right pieces for your project in the shortest amount of time? First you need to understand which pieces you don’t need, this will help you categorize and narrow in on the pieces that you do need.
To build the right combination for your project, questions like these are likely to come up:
- Which Service Mesh should I choose?
- How about API gateway?
- What about implementing a CI/CD pipeline?
- What’s the best way to deploy my database?
- What’s all of this going to cost me?
- How do I know which network policy will grant me the control and security I need?
The list goes on and we wanted the answers!
Q: What is the Main Lesson You Learned at KubeCon?
The evolution of Kubernetes is quite similar to the one that the cloud has gone through.
At first, we saw customers migrating their application as is with the not so optimal “lift and shift” method.
- Moving a complete web server directly to an EC2 instance
- Running their database on EC2 instances
After a while, companies realized that the cloud and particularly AWS could give them more than just a replacement for their servers. We started seeing them become more interested in managed services, more adapted services for DevOps, databases, security, storage and networking, and even serverless services.
The same happened to Kubernetes. The first migration was quick and dirty – just had to make it work! But through trial and error and years of experience we’ve learned that there is a plethora of features to tap into and some are more necessary than others in solving our project’s puzzle.
So our main takeaway from the event was that with Kubernetes there is no right and wrong answer, for every puzzle there is an optimal or most efficient solution, through hands-on experience you learn the best practices for your situation. And it’s a good idea to document these practices along your Kube journey. As a part of our Kubernetes ‘Domain of Expertise’ we are staying up-to-date on all the latest services K8 has to offer.
Thanks for the valuable insight Yossi and Tal. According to the questions you went there with it is clear that Kubernetes is a vast and evolving platform. Based on proven and well-documented best practices around containers and DevOps, it sounds like our team is equipped to help customers conquer their Kubernetes challenge. Puzzle solved.