Continuous Delivery Summit NA 2019

With tasks and pipelines, the Tekton pipelines project has powerful primitives for building a flexible CI/CD system. Event based triggering i.e. launching tasks and pipelines based on external events such as Github pull requests has been a much requested feature for Tekton.

In this talk you will see how you can use the Tekton triggers project to extract information from external events and use them in pipelines. But that's not all, you'll see how this approach can be easily extended to create any Kubernetes resources in response to any external event

Hayley will give a quick overview of the benefits of bringing security into your CI/CD pipeline. If security is not part of you CI/CD pipeline, you are either going to skip it, or it will bottle neck your process, undercutting the benefits of CI/CD. She will discuss best practices for integrating security into the CI/CD pipeline and share some lessons learned from her team at Snyk and the communities and clients with whom she works.

Jenkins on Kubernetes is typically a multi-tenant service where security is very important. In addition to multi-tenancy, when running CD workloads on a platform like Kubernetes there are typically other workloads deployed and if any workload does not have proper security configured it can impact all of the workloads running on the cluster.

The combination of Pod Security Policies (PSPs) with Kubernetes RBAC, namespaces and workload specific node pools allows for the granular security you need to ensure there are adequate safeguards in place to reduce the risk of unintentional (and intentional) actions that put your cluster at risk. Combining Pod Security Policies with other Kubernetes best practices allows you to provide the flexibility needed by CI/CD users while providing adequate guard rails so they don’t negatively impact other important Kubernetes workloads.

This talk will provide details on how to most effectively enable and use Pod Security Policies for Jenkins on Kubernetes.

Conversations around the adoption of Continuous Delivery and Spinnaker often center around tools and technology, and not enough time is spent discussing the organizational and process changes needed to ensure success. Using Spinnaker has implications for how teams develop software, organize themselves, implement compliance controls, and the traditional division of responsibilities between software, DevOps, and quality engineers.

In this presentation, I will explore these implications, and share the challenges I have encountered while implementing Continuous Delivery with Spinnaker. I will share how I went about solving these problems, my successes and failures.

There are no universal answers to these problems, but I will use my experiences as examples to provide insights to others that they can apply to their own efforts to increase engineering efficiency and deployment velocity/safety with Spinnaker.

Continuous Delivery on top of Kubernetes is changing the way that we build software, it comes with a great cultural change and challenges but a lot of advantages. While using tools like Jenkins X is an amazing starting point to get familiar with CI/CD and GitOps, companies can integrate with these tools in a whole new level to build the next generation Cloud Platforms. In this presentation, Mauricio will look at the Kubernetes resources provided by Jenkins X and how can we integrate our own domain specific concepts to leverage Jenkins X from inside our domain specific platforms. The presentation includes a live demo running on a remote Kubernetes Cluster and all the instructions for attendees to run the same example in their own environments.

Machine learning Ops is a practice similar to DevOps but focusing on automation for ML workflows. Kubeflow is an open source Kubernetes platform that allows you to run these workflows. Earlier this year a member of the Kubeflow community proposed using Tekton as the underlying workflow engine. This talk introduces the fundamentals of machine learning workloads and discusses the features we will need to add to Tekton to better support MLOps.