May 2020 Summaries

In this episode of TestGuild Automation Podcast, we dive deep into the world of AI-driven test automation with Oren Rubin, CEO and founder of Testim.io. We discuss how AI can help in generating end to end tests automatically from UI interactions, how it can improve the robustness of your tests by reducing false positives, and how it can handle changes such as replacing div tags with web components or custom elements. Oren also shares his thoughts on techniques and technologies that he wishes more front-end developers would use, such as writing more unit tests for front-end code and using Cucumber for BDD. Finally, we explore the new free playground for Playwright and Puppeteer users released by Testim.io.

In a recent webinar on The Future of Feature Management, LaunchDarkly announced early access sign-up for its Integration Framework and Feature Workflows. The Integration Framework enables developers to build custom integrations with other tools in hours instead of weeks. Partners can also integrate their solutions with LaunchDarkly and have them listed publicly. The framework allows users to send audit log events, store custom data on feature flags, and prevent flag changes until approval is met. Future plans include opening up the repository for more community participation and developing additional Integration Framework capabilities such as Metric Checks and Kill Switches. To join the Early Access Program, sign up today.

1. The myth of the bad apple refers to the belief that a single individual is responsible for causing problems in an organization, rather than acknowledging that complex systems are faultless and require adaptive capacity to handle incidents. 2. Continuous verification (CV) is a proactive experimentation tool that verifies system behaviors, as opposed to reactive testing methodologies that validate known properties. 3. Documenting best practices and runbooks can be valuable for communication purposes, but they do not improve reliability in complex systems because they cannot provide enough context or skill to improvise effectively during an incident. 4. The definition of resilience is having an adaptive capacity to handle incidents, which requires a level of improvisation and human adaptation that automated systems currently cannot replicate. 5. Self-healing algorithms can raise the floor of robustness in software systems but do not improve resilience because they rely on predetermined conditions and fail to account for the complexity of interactions within complex systems. 6. Feature flags are a powerful tool for managing dependencies and improving feature velocity, as they allow developers to make design decisions and change their minds quickly without causing major disruptions in production environments. 7. Effective incident response management involves studying resilience engineering principles, such as blameless postmortems or learning reviews, and understanding that complex systems require different approaches for navigating dependencies and managing risk. 8. Chaos engineering can be a valuable practice for improving system reliability by encouraging teams to discuss potential failure points and run experiments in staging environments before moving them into production.

In this episode of Testing in Production, Yoz Grahame is joined by Dylan Etkin, CEO of sleuth.io, to discuss measuring impact to changes in production. They talk about SLIs and deriving new SLIs, especially working out the value of them. They also discuss how to derive correlations between rise in one stat versus SLO problems or if it's going to end up causing more false positives than you think. Finally, they talk about using user behavior for SLIs and SLOs or is it just too much noise?

LaunchDarkly has been named a Best Place to Work in the San Francisco Bay Area. The company prioritizes culture and values inclusivity, transparency, diversity, and active listening. Despite recent events, LaunchDarkly continues to grow and improve its platform, which enables customers to deliver features safely and controlled. The recognition is appreciated, but the company acknowledges that continuous improvement is essential as it grows.

In this episode of Test in Production, we discuss Real User Monitoring (RUM) and Synthetic Monitoring with Zach Henderson from Catchpoint. We explore how to use these monitoring techniques to improve the performance and reliability of your applications. We also delve into some interesting bugs that can arise when using these tools. Some key takeaways include: - Real User Monitoring (RUM) involves collecting data about how real users are interacting with your application, including metrics such as page load times, error rates, and user engagement. - Synthetic Monitoring involves simulating user interactions with your application to proactively identify performance issues before they impact real users. - To effectively use RUM and Synthetic Monitoring, it's important to establish clear goals and metrics for success, as well as a plan for how you will respond to any issues that are identified. - One interesting bug that can arise when using these tools is the issue of network time synchronization (NTP). This can cause delays in user interactions with your application if not properly addressed.

LaunchDarkly has made significant progress in 2019, serving over 2 trillion flags per day across thousands of customers and introducing product updates such as Experimentation and Data Export. The company's new feature, Feature Workflows, enhances its core platform by allowing users to create complex workflows within LaunchDarkly that better map to how their teams build, ship, and control software. Key components of this update include Scheduling, Approvals, and Metric Checks, which enable users to pre-define flag changes for future points in time, request or require approvals before making flag changes, and trigger dynamic changes to flags based on real-time operational and business metrics. Feature Workflow Templates allow teams to create standardized release processes within an organization by combining these components together. The update aims to ensure consistent best practices across teams, automate and safeguard changes to features, and create standardized release processes within an organization.

This week's discussion focused on deployment strategies and how they have evolved over time. The text highlights the importance of knowledge sharing and learning from past experiences to improve processes. Deployment strategies vary among companies, with some using continuous delivery while others rely on manual deployments. As technology advances, it is crucial to ensure that software practices remain approachable and enable people rather than hinder them. Sharing knowledge within the DevOps community helps create better processes and contributes to industry growth.