December 2018 Summaries

During the AWS re:Invent 2018 event, Dynatrace served a significant amount of beer and wine, which highlighted the challenge of scaling services to meet user demands, akin to the challenges faced by App Dev, Operations, and Business teams. To address these challenges, Dynatrace has embraced an Agile approach and a culture of experimentation, breaking down monolithic systems into scalable microservices. This transformation has resulted in substantial performance improvements and notable statistics, such as achieving 26 major releases annually and reducing bugs in production by 93%. The company's journey from Waterfall to DevOps and NoOps, along with their agility, speed, quality, and availability, are further detailed in a Forrester Case Study. The event also provided a dataset for analysis, thanks to the participation of over 1,200 attendees, with Dynatrace aiming to continue engaging with the community at future AWS events.

The text discusses the adoption of agile methodologies and cloud-native platforms by companies to facilitate faster application delivery and responsiveness to customer needs, emphasizing the necessity of a Continuous Integration and Continuous Delivery (CI/CD) pipeline to automate the building, testing, and deployment processes. A significant challenge addressed is transitioning services from testing to production without affecting uptime, for which the blue-green deployment strategy is highlighted as a solution. This approach involves maintaining two identical production environments, enabling seamless transitions between them. The blog post uses Cloud Foundry and Dynatrace to demonstrate this strategy with the Dynatrace Sockshop application, a microservices demo comprising eight services, each with its CI/CD pipeline managed by Concourse. The text details the steps involved in the CI/CD pipeline, from building in the dev environment to a blue-green production deployment, while Dynatrace is utilized for monitoring and performance analysis to ensure smooth transitions and identify any issues. Although Concourse is presented as effective for continuous integration, Spinnaker is suggested as a more suitable tool for continuous delivery, with plans to explore this combination in future discussions. The text concludes by inviting readers to try Dynatrace for monitoring their services in Cloud Foundry, offering guidance on setup and best practices.

Migrating applications to the cloud often involves using the 6-R Migration Patterns, with monolithic applications frequently undergoing re-architecture to break them into smaller components. This process can be facilitated using tools like Dynatrace, which provides insights into application dependencies and performance without requiring code changes. The approach was presented in a blog series and showcased through a demo at AWS re:Invent 2018, where a sample application, easyTravel, was containerized and deployed on AWS Fargate, demonstrating improved performance over its original monolithic form. The demo emphasized the importance of understanding application architecture, identifying key endpoints and processes, and using tools like Dynatrace for monitoring and optimization throughout the migration process. The method encourages refining extraction points and comparing performance between monolithic and containerized versions to ensure successful migration and optimization.

Site Reliability Engineering (SRE) is a critical discipline focusing on building reliable systems, but the concept of "self-healing" in this context can be misleading, as it's more about automating remediation actions than systems independently healing themselves. The author suggests using terms like Smart- or Auto-Remediation and emphasizes that self-healing should not be limited to operations in production but should also involve proactive measures to prevent issues from reaching production, a concept they call "Shift-Left SRE." This involves integrating quality gate checks into the CI/CD pipeline to detect and mitigate potential problems in pre-production environments, thereby enhancing system resilience. Using tools like Dynatrace, auto-remediation as code can be implemented, where automated scripts focus on fixing root causes rather than symptoms, creating a more efficient incident response. The ultimate goal is to establish an "Unbreakable Delivery Pipeline" by combining automated validation and deployment of remediation scripts, thus minimizing production issues and fostering collaboration between operations and engineering teams for better system resilience.