Continous Delivery at Scale event introduction

As Fred Brooks so famously said in "The Mythical Man Month" "You can't make a baby in a month with 9 women" so how do we scale software development?
What does it take to build complex systems on a massive scale? What should we scale and what shouldn't we?
Spoiler, the answer is not bigger, more complex procedures and processes, stand aside SAFe! This stuff is about applying information theory and engineering principles to software development.

We broke monoliths into microservices and leverage events to orchestrate business processes across all moving pieces. Why are not applying the same event-driven concept to delivery and operations automation?
This session introduces the CNCF project Keptn (www.keptn.sh): An Event-Driven Control Plane for automating Continuous Delivery, SLO-Based Quality Gates, SRE Automation and Auto-Remediation.
Besides an intro to event-driven CD & CO you will see live demos + how the community is adopting Keptn as well as how you can extend and contribute

Who wants to maintain a large, complex and ever-growing continuous integration and delivery pipeline? 
In this presentation, we introduce Tekton, an open-source platform for creating cloud-native CI/CD pipelines. We present how to use Tekton to break down complex pipelines, reducing and distributing their maintenance cost. We will also share lesson-learnt stories from running Tekton at scale.

The usual method of delivering software into production consists of developers creating the artifact, then handing it off to a team of DevOps engineers who sweat bullets as they write scripts and carefully deploy code into production. This method is time-consuming, error-prone, and is no longer suitable for the accelerated pace of business.
In this talk, Ravi will discuss the need to combine the practice of Continuous Delivery with Machine Learning to give power directly to the engineers - including actually how to do it.

The talk mentions the current problems of GitOps, moves on to a vision of GitOps 2.0 and finishes with the announcement of the GItOps working group by Amazon, Codefresh, Github, Microsoft and weaveworks.

There is an almost bewildering amount of machine learning platforms available today, both open source and commercial. Some focus exclusively on the development of models without much regard for a model's entire lifecycle, while others assume a trained or tuned model already exists. End-to-end platforms that support development and deployment of models and aid users in going from prototype to production are a fairly recent phenomenon that is nevertheless not as rare as one might think.
On this safari, we shall encounter different platforms spotted in the wild at companies across the globe, observe their characteristics from a safe distance, and attempt a taxonomy based on the underlying infrastructure. 
Fun for the entire family!

A free-form discussion on Continuous Delivery challenges and your chance to participate in round-table discussion

Learn how to scale public cloud to your on premise clusters with CI/CD MLOps pipelines from cloud on edge
We will show how you can detect model drift and then use events to trigger a new job to train on new data and deploy a new model using Azure MLops in no time.

In FB Lite world, we have fully automated server releases without humans taking a single decision or doing any manual work. We will talk about how the whole flow is running by covering detection, attribution and remediation steps.

Learn how Microsoft scaled its build and test infrastructure to meet the demands of the largest repos in the world while also enabling a performant retail desktop experience for developers.

Learn how Quora leverages post-push review, distributed batch testing, and automatic reverts to deploy dozens of times a day to a monolithic web application.

Among the Netflix client application's top operational risks is the deployment process. Unlike the server applications, client applications run on systems (PCs, TVs, phones) we don’t control which results in unique challenges. This means there are opportunities to further reduce friction, improve velocity, and the release and rollback decision. In this presentation, we will talk about how we deliver, detect and orchestrate the release of client applications.

Over 900 engineers at Datadog do thousands of deployments per day, to hundreds of services in different environments, regions, and cloud providers. Embracing continuous delivery and automating operations is critical to scaling our teams.
To accommodate the diversity of the applications our engineers deploy, we needed to build a software delivery platform that could be extended to meet application-specific use cases, especially for teams deploying stateful applications.
In this talk we will discuss the primitives our software delivery platform exposes, the design decisions we made, and how teams extend the platform for their particular use cases. To illustrate, we will explore a case study of automating rollouts and common operational workflows of Apache Kafka on Kubernetes.

A free-form discussion on Continuous Delivery challenges and your chance to participate in round-table discussion