Carlos Sanchez

Moving to Kubernetes opens the door to a world of possibilities, but as the number of resources, clusters and namespaces grows, it becomes increasingly difficult to manage.

We will show how at Adobe Experience Manager we continuously deliver changes to cluster resources and have implemented a way of doing feature flags and progressive rollouts with Kubernetes yaml definitions, across multiple clusters and regions using Jenkins, ArgoCD, and other open source projects.

Running a cloud service in multiple Kubernetes clusters across regions, forced us to have high levels of automation, tied with monitoring and alerting across the globe. The typical release train model (dev-stage-production) had shortcomings that soon became evident. Adopting Progressive Delivery allows us to move faster, safer, and reduce the mean time to recovery in case there are any issues. Both canary deployments and feature flags are typically used to implement Progressive Delivery at both application and operational service levels.

Teams collaborate with a DevOps philosophy and a "you build it, you run it" approach, with freedom, but also requiring tooling to help teams deliver value with the least interference as possible. We are continuously improving this tooling to make it as easy as possible to safely deploy changes.

Progressive Delivery makes it easier to adopt Continuous Delivery, by deploying new versions to a subset of users and evaluating their correctness and performance before rolling them to the totality of the users, and rolled back if not matching some key metrics. Canary deployments is one of the techniques in Progressive Delivery, used in companies like Facebook to roll out new versions gradually. But good news! you don't need to be Facebook to take advantage of it.

We will demo how to create a fully automated Progressive Delivery pipeline with Canary deployments and rollbacks in Kubernetes using Jenkins X, an open source platform for cloud native CI/CD in Kubernetes, and Flagger, a project that uses Istio and Prometheus to automate Canary rollouts and rollbacks.

Building and testing is a great use case for containers, both due to the dynamic and isolation aspects, but running in just one machine is not enough and quickly needs to scale to a clustered setup. But which cluster technology should be used? Docker Swarm? Apache Mesos? Kubernetes? how do they compare? All of them can be used to dynamically run a cluster of containers.

Building and testing is a great use case for containers, both due to the dynamic and isolation aspects, but running in just one machine is not enough and quickly needs to scale to a clustered setup. But which cluster technology should be used? Docker Swarm? Apache Mesos? Kubernetes? how do they compare? All of them can be used to dynamically run a cluster of containers.

The Jenkins platform is an example of dynamically scaling by using several Docker cluster and orchestration platforms, using containers to run build agents and jobs, and also isolate job execution.

This talk will cover these main container clusters, outlining the pros and cons, the current state of the art of the technologies and Jenkins support.

The presentation will allow a better understanding of using Docker in the main Docker cluster/orchestration platforms out there (Docker Swarm, Apache Mesos, Kubernetes), sharing my experience and helping people decide which one to use, going through Jenkins examples and current support.