Cloud Native Computing Foundation KubeCon + CloudNativeCon North America 2022, 11 Nov 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keynote: Beyond Automation: Kubernetes Success Requires a GitOps Mindset - Shatarupa Nandi

Description

Don’t miss out! Join us at our upcoming event: KubeCon + CloudNativeCon Europe in Amsterdam, The Netherlands from April 17-21, 2023. Learn more at https://kubecon.io The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.

Keynote: Beyond Automation: Kubernetes Success Requires a GitOps Mindset - Shatarupa Nandi, Senior Director of Engineering, VMware Tanzu

Speakers: Shatarupa Nandi
As an industry we’ve come a long way to embrace automation. What once seemed like a scary proposition is now at the heart of a successful Kubernetes platform, and as we evolve as practitioners we’ve come to embrace GitOps to fully realize the value we can provide: value to the developer experience, security, and operations at scale across multiple environments.

In this talk, VMware's Shatarupa Nandi, will discuss how we can ship high quality software securely and repeatably on Kubernetes, and still keep developer experience simple. We'll discuss how the IT industry has embraced GitOps for applications, for 3rd party software, and for Kubernetes itself. We'll talk about how this mindset gives businesses a competitive edge and helps showcase IT’s strategic value and importance.

A

Hi everyone I'm chaparrupa I work on Project carvo, where our goal is to provide a set of Reliable Tools that help you automate and run your software in production, and today, I want to share a real world story about a platform operations team. That's been able to manage over 30 clusters and several applications in production in a highly regulated environment using the carpool tool chain.

A

But before that, I cannot contain my excitement and I have to share that project. Carvel was just accepted as a Sandbox project in the cncf and I am so happy to be here today saying that.

A

Okay, now back to the story, um so let me talk a little bit about the challenges that this team faced on their production Journey. So, first, their applications have complex deployment topologies. So installing these applications by following a Playbook of scripts was toil heavy. It was error prone and their development teams really didn't like that experience.

A

They, the development teams, didn't have the right level of access to run these scripts. They didn't know where to run the scripts from and managing fine-grained access was hard for the platform operations team.

A

As all a lot of us today do they also depended on a lot of Open Source software on third-party software, and they needed to have a way to ensure that the software was built securely. It was easily accessible in their various environments and it was easily customizable security. Compliance reliability is really important to this team, so this team knew that the least error prone and the most user-friendly way to satisfy these criteria was to do so programmatically.

A

They knew they wanted to adopt the get-offs mindset for managing their clusters and what really resonated with them about the get-offs mindset was this idea of deploying continuously declaratively and doing so collaboratively as an organization.

A

So what did they do? The team first started to keep configuration in the central git repository. This repository became a central point of collaboration for the entire organization. They could Source changes to this repo from various teams.

A

Then they started relying on Carvel's package manager, cap controller, to create clusters. Following this configuration now they could make a single change in their configuration and Cat controller would automatically roll it out to their many clusters.

A

Previously, updates of this kind took days of coordinative coordinated effort across multiple development teams, but now they were able to update their applications and their clusters in Dev, test and prod. In minutes they could also automatically provision production like clusters in Dev, which was really helpful for their testing.

A

The next key thing for them was continuous reconciliation, so Carvel's cap controller provided much better accountability. When rolling out centralized configuration, then a Playbook driven approach used to provide for their development teams using cap controller, they were guaranteed that the cluster converged with the configuration every 10 minutes.

A

This prevented configuration drift and really eliminated the problem of snowflake clusters. For them, it also meant that they could guarantee that their deployments were compliant with the declared policies.

A

Then they started bundling their applications configuration its kubernetes manifests and its dependencies in a single immutable, oci artifact using a Carvel tool called image package.

A

This oci artifact could now be signed, it could be relocated and it could be referenced by its unique Digest and that's how they are today able to relocate their apps and their dependencies to clusters, including to their Edge clusters.

A

And finally, they were able to use Carvel's, yaml wrangling tool, ytt to write overlays for third-party software. The overlays could be checked into git and they were integrated into their git Ops deployment system that that was powered by cap controller, as we talked about earlier.

A

So this empowered their developers to leverage that centralized configuration, but also provided them. The flexibility they needed through environment-specific overlays.

A

So with this, the platform operations team were really able to enable the development team to provision new clusters with common software. You know like cert manager, Prometheus fluent with external DNS and other commonly used things with, but with approved versions, approved configurations in a matter of minutes, and they were also able to keep these clusters upgraded.

A

They upgrade multiple times a month to you, know, stay on top of cves and so on, and it's been awesome for us on Project Carvel to support this platform operations team to have them be part of the community to co-develop with them and they've been a big part of our journey to cncf.

A

So if you want to manage your production software on kubernetes securely reliably and with automation, built on top of native kubernetes apis following get-offs principles, then come join us on Project Carvel. You can find us on carvel.dev. You can find us on the kubernetes slack we're super eager to build in the open with you super eager to see you leverage it in various cncf projects as building blocks for your own projects. The team and I will be at the VMware Booth later today tomorrow and we'll have lots of talks and demos and so on.

A

There so hope to see you. There hope you have a fun coupon and hopefully continue this conversation. Thank you. Everyone.