CNCF-hosted Co-located Events North America 2024

In this presentation we will share our Continuous Integrations (CI) pipeline extensions via Argo Workflows in a pure cloud-native way seamlessly integrated with our code versioning system. We will also describe a large-scale asynchronous event-based bus communication that has accelerated our CI pipelines but also increased their resiliency. Finally, we will cover some Continuous Delivery enhancements via ArgoCD and Argo Rollouts. Actually, even our own infrastructure is operated via GitOps model. This collection of open-source projects and internal tools provides a full cloud-native orchestration and gitops experience for Dev, QA and Performance teams. Our solution extends the existing GitLab CI/CD capabilities by providing the ability to orchestrate, build and manage service interdependencies and integrations with other systems and enhanced metrics-based service rollouts in a multi-region cloud architecture.

Since the day Kargo was released, I have been exploring the idea of using it not only to deliver and promote applications but also to deliver infrastructure through its progressive delivery capabilities. Using Kubernetes-based tools like Crossplane or Pulumi, we can define infrastructure as code and deliver it progressively to our management clusters and then promote this infrastructure through different stages without the need for extra CD script magic. Let me show you how Kargo helps platform engineering streamline and automate the progressive rollout of infrastructure changes to all stages. This talk will cover the basics of Kargo and how to use it with Infrastructure as Code tools.

GitOps is an industry standard best practice in the Kubernetes space. But there are still gaps in the developer experience. Change previews and environment promotion have been two of the main pain points. Upcoming Argo CD features fill this gap by providing automated change previews as PR comments and by managing environment promotion by automatically opening and merging pull requests. This talk presents a live demo of these new Argo CD features. We’ll show you how change previews will help your developers merge changes faster and with more confidence; we’ll show how you can design a change promotion strategy tailored to your organization’s needs; and we’ll show how your developers can monitor change promotions via the Argo CD UI. Promotion-by-PR is one of several options for managing GitOps promotions. Besides showing a demo, we’ll compare and contrast this strategy with other open source and vendor-based solutions so that the audience can select the system that matches their needs.

Argo Rollouts makes Progressive Delivery easy to adopt, but some times things do not work as expected. Are the steps correctly set? are the analysis metrics right for the workloads? At Adobe Experience Manager we deploy over 10k customer services to Kubernetes. Changes can occur multiple times per day both internal and from code. A new feature can work fine for 99% of customers but still affect the other 1%, and detecting this just from tests is costly. Enter Argo Rollouts, which allows deploying new versions to a subset of users before rolling them to the totality of the users, and rolling them back if not matching some key metrics, using techniques like canary deployments. We will show our learnings deploying Argo Rollouts to manage over 10k workloads using canaries, how do we balance speed and safety for our customers, and some of the issues that we have faced when adopting it.

Are you tired of looking at metrics and logs after each deployment? Do you learn about failed deployments from unhappy customers? Did you always want to deploy on Friday afternoon and go straight to the pub? Many teams perform “blind” deployments without any real insight into what will be affected by the new application version. Consequently, they don’t have enough data to understand the blast radius of a release and whether to decide if it was successful or not. Even companies that have several metrics in place, don’t always use them in an automated manner. Wouldn’t it be great if you could see user behavior with new features in real time and identify performance bottlenecks before a full release?

In this talk, we will focus on common scenarios regarding Argo Rollouts and observability metrics, we will explain:

• Minimum requirements in terms of tools and metrics/traces/logs 
• Well-known observability use cases
• Common automation pitfalls
• RED/USE metrics tradeoffs