May 2021 Summaries

Server-side testing is a technique that expands upon client-side testing, allowing for more sophisticated tests on deeper parts of applications beyond just front-end features. It renders all tests on the server instead of the visitor's browser and has gained traction in various industries, particularly among ecommerce retailers. The key differences between client-side and server-side testing include the ability to test complex back-end elements with server-side testing, as well as its capacity for deep insights into full application stacks. Common methodologies for server-side testing include A/B and multivariate testing. Feature flags are a valuable tool in running server-side tests, allowing developers and product managers to toggle feature functionality on and off without deploying code. LaunchDarkly's feature management platform offers an experimentation add-on that enables both front-end and back-end (server-side) experiments.

Mark Burry, Senior Developer at iPipeline, discussed their journey with feature flags and how they eventually found fitness functions and flagging conventions as best practices during a talk at LaunchDarkly's user conference, Galaxy. The company had issues with naming conventions, feature flag management, and tech debt, which led them to implement fitness functions and processes to ensure that their flags were well maintained and loved throughout their life cycle. They focused on code references, deploying a tagging strategy, running automated assessments on their flags, and implementing web hooks for continued visibility.

A canary release is a software deployment strategy where new features are made available to a limited group of users before being rolled out to the entire user base. This approach helps improve software quality and reduces risks associated with shipping updates. The name "canary" comes from an old mining practice where miners would use canaries to detect toxic gases in coal mines. Canary releases allow developers to identify bugs, breaks, and hazards of new features before a wider rollout. While similar to blue-green deployments, canary releases are more targeted towards specific user groups. Feature flags can be used alongside canary releases to simplify the release process by introducing new code to a small portion of users within the main production environment.

Cloud deployment models define where servers are located and who manages them. These models include public clouds (owned by a third-party provider), private clouds (dedicated to one user or organization), hybrid clouds (combination of public and private clouds), multi-clouds (use of multiple cloud providers), and community clouds (for specific communities). Each model has its advantages and disadvantages, depending on factors such as cost, scalability, security, and compliance requirements. Additionally, there are three main service models: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS), which determine the level of control and management provided by cloud providers.

Deployment automation is an approach that allows organizations to increase their velocity by automating build processes, testing, and deployment workflows for developers. This enables them to release new features faster and more frequently. Deployment automation focuses on deploying software automatically to different environments, running checks against the software, and pushing the software build to production with limited or no human intervention. It helps detect bugs much faster, saves time due to a faster feedback model, and improves accuracy by automating tasks such as environment setup, testing, and database provisioning. Key best practices include continuous testing, configuration management for environments, and monitoring and logging. Benefits of deployment automation include reduced lead time, faster bug detection, avoidance of repetitive tasks, and increased accuracy. Deployment automation can be combined with continuous delivery or continuous deployment to further enhance the release process.

Release testing is a crucial coding practice and test strategy aimed at ensuring the stability and functionality of software release candidates before they are deployed to users. It involves various methods such as unit testing, end-to-end testing, smoke testing, regression testing, load testing, and acceptance testing. The goal is to identify and eliminate errors and bugs in the software release candidate, thereby building confidence in its readiness for actual use. With the advent of continuous integration and continuous delivery (CI/CD) practices, modern testing processes have evolved to be more efficient and less risky, allowing developers to test new functionality on live production servers using feature flags.

Automation and orchestration are distinct yet interconnected concepts in the realm of IT operations. Automation refers to executing a specific task without human intervention, while orchestration involves managing multiple automated tasks to create a single workflow. Both automation and orchestration are crucial for improving productivity, reducing costs, and streamlining processes in complex IT environments. Automation can be achieved through various means, such as scripts or tools, and is often used for infrastructure, software, deployment, and business process tasks. Orchestration, on the other hand, requires a platform to manage automation effectively and make decisions based on events. Examples of orchestration platforms include container orchestrators and CI/CD systems. In DevOps culture, both automation and orchestration play vital roles in shortening the system development life cycle. LaunchDarkly's feature management platform offers built-in automation for better managing feature releases and integrates with various tools to facilitate better orchestration and automation. By understanding and leveraging these concepts, organizations can streamline their infrastructure and software development processes effectively.