Les process CI/CD sont souvent chronophages, coûteux en ressources machine, et difficiles à orchestrer avec précision. Découvrez comment avec un langage d'orchestration il est possible d'optimiser grandement certains processus, mutualiser les étapes et ressources communes, définir avec un haut niveau de granularité les éléments nécessaires et l'ordonnancement des actions. La présentation est réalisée en trois grandes étapes : - La problématique des CI/CD : en quoi sont-elles peu efficaces, les limites de leur expressivité, et leurs marges d'optimisation ; - Le langage Mélodium : pourquoi un langage de programmation, l'avantage d'une expression de l'orchestration, et les caractéristiques d'un moteur à programme distribué ; - L'application d'un langage orienté orchestration à la CI/CD : intégration dans une CI existante (Github Action, avec portage automatique), les étapes d'amélioration usuelles après un portage et patrons de conception possibles, utilisation dans une infra, optimisations et réductions de consommation associés.

Automating Deployment with GitHub Actions: Learn how to use Terraform in GitHub Actions for continuous integration and deployment.

Using Terraform in GitHub Actions to automate deployments.

Using Terraform within GitHub Actions to automate deployments and CI/CD workflows.

This presentation highlights practical validation techniques to prevent misconfigurations and enhance reliability in Apache Airflow environments. We cover two key safeguards: validating that sensors are correctly tied to their upstream tasks, and checking that critical DAGs have PagerDuty alerting enabled. Both validations are automated and integrated into our CI/CD pipeline using GitHub Actions, ensuring continuous enforcement and early detection of potential issues before deployment. In addition, we’ve implemented a solution to track Service Level Objectives (SLOs) for our DAGs, enabling better insight into reliability and performance over time. These checks form a practical defense against operational blind spots, promoting workflow reliability and robust incident response. Join us as we uncover practical strategies to streamline workflow monitoring and enhance system resilience using Apache Airflow's robust capabilities.

Second workshop in our series on financial data automation and analysis. Outline covers Data Acquisition & Preprocessing (Yahoo Finance API, SQLite storage, generating additional variables); Exploratory Analysis & Time-Series Features; Forecasting Techniques (Prophet, ARIMA, Deep Learning); Visualization & Deployment (Plotly, Streamlit Cloud, GitHub Actions).

I am a big believer in synthetic monitoring as a concept to build reliable production services. If engineers are supposed to run what they build, they need monitoring tools that work for them. I have built my own custom solutions in the past using Jenkins or GH Actions and later used SaaS tools for this. I want to share my experience how I got frontend engineers to build monitoring and get everyone on an engineering team to care about production system reliability. Daniel Paulus is an accomplished technology leader, presently leading as the VP of Engineering at Checkly, building synthetic monitoring with Playwright.

Learn the fundamentals of DevOps best practices and how to continuously deploy cloud resources. Walk through configuring Pulumi GitHub Actions to deploy AWS resources programmatically.

Learn the fundamentals of DevOps best practices. You will become familiar with the core concepts needed to deploy cloud resources continuously. Walk through configuring Pulumi GitHub Actions to deploy AWS resources programmatically and accelerate your cloud projects with the skeleton code provided.

In the dynamic sphere of cloud-native development, constructing robust, secure, and efficient CI/CD systems is not just desirable but essential. This talk will take you on a journey through the modernization of our build pipeline and CI/CD systems, specifically focusing on our project, Parca - a performance analysis tool that operates at the kernel level using eBPF technology. We will explore the challenges of ensuring security and compatibility across various kernel versions, a critical requirement for kernel-level operations like Parca. Our solution is a fully open-source, cross-platform CI/CD pipeline, uniquely tailored to meet Parca's intricate needs. This pipeline exemplifies the power of modern tooling, stepping beyond the limitations of traditional makefiles. Critical aspects of our discussion will include: Byte-by-Byte Reproducible Builds: Delving into how we achieve exact replicability in our builds to enhance security and fortify the supply chain against vulnerabilities. Locally Reproducible CI/CD: Demonstrating the methodologies that ensure what is run and tested locally can be precisely reproduced on remote CI servers, bridging the gap between local development and production environments. Modern Tooling for CI/CD Pipelines: Showcasing how tools like Jetpack Devbox (easy way to use NixOs), Dagger, Mage, GitHub Actions, and Zig's build system have revolutionized our approach to building CI/CD pipelines. Collaborative Efforts and Open-Source Innovations: Highlighting the cooperative nature of our project and how open-source tools and community contributions have been instrumental in our success. Attendees will gain valuable insights into creating a secure, adaptable, and efficient pipeline in an open-source environment. This talk will detail our journey and solutions and provide practical lessons for those looking to modernize their CI/CD systems in the cloud-native ecosystem. Join us to explore the frontiers of CI/CD modernization and learn how your projects can benefit from these cutting-edge practices and tools.

Learn the fundamentals of DevOps best practices. You will become familiar with the core concepts needed to deploy cloud resources continuously. Walk through configuring Pulumi GitHub Actions to deploy AWS resources programmatically and accelerate your cloud projects with the skeleton code provided.