DevOps

While writing a Dockerfile is easier than ever (AI can help you generate a sample in a second!), but the resulting images might not be good enough. In this talk, we will discover a comprehensive guide to optimizing Docker images for JVM applications. Learn a series of tactics to reduce image size, speed up build time, and enhance overall performance and developer experience. We will illustrate the performance improvements achieved by applying these tactics to commonly used web frameworks. Additionally, we will explore the use of Gradle plugins to streamline the optimization process.

What we learned from our previous vote4.hk project, and how we design wars.vote4.hk to optimize cost and workflow of an opensource project.

The workshop will give a hand-on experience about Kubernetes and also how to manage the Kubernetes cluster in GitOps via Weaveworks FluxCD

By using Ansible for DevOps, we could manage both VM, Docker image provision, Kubernetes and CephFS provision, or even Kubernetes Pod runtime management.

GitHub Actions is now the second most popular Build Platform on GitHub.

But it is also much more than than a CI system: it is a very powerful automation engine.

Let take a look at how we can do CI with Actions, and how we can automate operations on our GitHub projects.

As a great glue language, Python could (definitely) orchestrate many modern and powerful tools to build a scientific cluster. This talk will illustrate one of the real world examples, which is a cluster built and running on top of LXD/LXC/KVM/MaaS/public clouds so you may thus be a amateur CFD (computational fluid dynamics) or physics researcher.

Starting with Ansible installation, playing with CLI, writing Ansible Role with test on LXD, create Docker Image with Ansible Playbook and Roles, and managing a running Kubernetes with Ansible.