November 2019 Summaries

In an episode of the podcast "Break Things on Purpose," hosts Rich Burroughs and Jacob Plicque interview Haley Tucker, a Senior Software Engineer at Netflix, to discuss Netflix's advanced Chaos Engineering practices. They explore how Netflix has evolved its Chaos Engineering from early tools like Chaos Monkey to more sophisticated platforms like the Chaos Automation Platform (CHAP) and Monocle. These tools allow Netflix to conduct complex experiments that combine chaos and load testing with canary analysis, aiming to optimize the resilience and performance of their services. Tucker explains that while Chaos Monkey was pivotal in improving system redundancy, the current focus is on targeted experiments that provide actionable insights for service owners. The conversation also touches on Netflix's approach to resilience, the role of human factors, and the importance of understanding system behavior under stress to maintain high standards of user experience and service reliability.

KubeCon + CloudNativeCon North America 2019 in San Diego gathered between 10,000 and 12,000 attendees, highlighting the mainstream adoption and vibrant community of Kubernetes. The event featured numerous talks and keynotes that explored new features like Ephemeral Containers for debugging, Open Policy Agent for managing Kubernetes policies, and the integration of OpenCensus and OpenTracing into OpenTelemetry for clearer tracing standards. Attendees were introduced to innovative projects like Vitess from PlanetScale, which enhances MySQL clustering, and clusterlint from Digital Ocean for pre-upgrade config checks. The conference emphasized user experience and community-building, with many sessions focusing on practical challenges and real-world applications, such as zero downtime deployments and Chaos Engineering. Social activities, including a block party and wellness events, fostered networking and relaxation, with attendees encouraged to balance session attendance with informal interactions. The conference left participants optimistic about the future of Kubernetes and inspired to explore new tools and methodologies.

AWS re:Invent is the largest cloud computing event worldwide, and the Gremlin team provides several engaging activities to attendees to enhance their experience, including a booth at the Expo, a Chaos Engineering-themed memory challenge, and a Reliability Tour with stops at various partner locations. The Reliability Tour offers participants a chance to win prizes by collecting enamel pins from different tour stops, including Gremlin's booth, MongoDB's Pub Crawl, and others. Attendees can also participate in hands-on workshops, such as running Chaos Engineering experiments on Amazon EKS, and attend reliability-focused talks. Additionally, there are opportunities for networking and socializing, such as a Pub Crawl event with MongoDB and Datadog. For those new to the event, resources like A Cloud Guru's guide can help navigate the overwhelming experience, and Gremlin encourages participants to start a free trial of its platform to find and fix availability risks proactively.

Gremlin has introduced native Kubernetes support to its Reliability as a Service platform, allowing users to easily discover, visualize, and target Kubernetes objects for chaos experiments. This integration simplifies the process of targeting containers within a Kubernetes cluster by abstracting the complexity of container management, making it easier to conduct experiments that enhance system resilience. Users can now target services running on Kubernetes objects, such as Deployments and Pods, through a web app that offers a visual map of the cluster. This development reduces the time spent on negative testing and allows for more efficient identification and testing of potential failure points within systems. The native integration is now available for all users, with further enhancements to Kubernetes targeting within scenarios planned. Gremlin's platform aims to empower users to identify and address availability risks proactively, offering resources like an ebook on Kubernetes reliability and a free trial to explore its features.

In a rapidly evolving software development landscape, traditional change management processes struggle to keep pace, prompting a shift towards Continuous Integration (CI), Continuous Delivery (CD), and Continuous Deployment (CD) methodologies. These CI/CD practices emphasize frequent, small code changes that are easier to integrate, test, and roll back, thereby enhancing software release speed and reliability. The integration of DevOps culture further bridges the gap between development and operations, fostering collaboration and automating processes to boost organizational efficiency. Despite fears about the implications of "continuous" methodologies, they are merely modern approaches aimed at accelerating deployment without compromising safety. The transition from monolithic architectures to microservices supports this shift, enabling isolated, scalable, and independently testable services. Automation and rigorous testing, including Chaos Engineering, are crucial to ensuring code stability and reliability. This approach is supported by research indicating that high-performing companies achieve greater operational efficiency and customer satisfaction by adopting lightweight change approval processes and automated testing. Ultimately, embracing CI/CD alongside Chaos Engineering and nurturing a collaborative DevOps culture allows organizations to safely and swiftly deliver high-quality software.

Gremlin offers a solution to help engineers prioritize reliability work by using their Reliability Calculator, which assists in deciding where to focus efforts across multiple services within an application's architecture. Tammy Butow, an expert with a decade of experience, advises prioritizing services that are critical and struggling but not beyond recovery, suggesting that improving a "B" service to an "A" yields more value than trying to salvage a "D" or "F" or trying to enhance an already "A" service. The Reliability Calculator allows users to input details about their services to receive a reliability grade and personalized recommendations for improvement, making it easier to plan reliability efforts effectively. Gremlin's platform is designed to help identify and fix availability risks before they affect users, offering a free 30-day trial for new users to explore hidden risks in their systems.