March 2018 Summaries

LaunchDarkly projects simplify managing feature flags across multiple software projects. With environments, users can easily manage their feature flags throughout the continuous delivery pipeline. However, as companies grow and accumulate more projects and environments, navigating between them becomes challenging. To address this issue, LaunchDarkly has introduced an improved environment switcher that allows users to find any environment quickly, regardless of the number of projects or environments. Additionally, favorites have been added, enabling users to mark frequently used environments for quick access.

In their first week at a new job, the author experienced series B funding and celebrated with a trip to a local tiki bar. They have been learning a new programming language, deployment toolchain, and getting acquainted with colleagues while realizing there's much more for them to learn. The author has previously used LaunchDarkly for feature management at their last job, where they experienced an operational issue due to a bug that created a channel in RabbitMQ every time an event was published. This led to an outage when the RabbitMQ server ran out of memory and prevented all writes. The author is now excited about using feature flags for operational scenarios at their new job, as it allows them to disable publishing events for specific customers during high traffic situations, preventing service degradation and cascading issues. They anticipate continued growth and learning in this role.

In this podcast episode, Ellen Chisa, John Kodumal, Edith Harbaugh, and Paul Biggar discuss their experiences as startup founders. They share stories about how they started their companies, the challenges they faced, and the lessons they learned along the way. The conversation covers topics such as fundraising, management practices, and the importance of having a clear vision for your company.

Jonathan has joined LaunchDarkly as VP of Engineering after spending 12 years at Atlassian where he worked with John Kodumal and many current LaunchDarkly team members. He was part of the growth of Atlassian from a small company to over 2,500 employees. Jonathan has experience in building feature management systems and believes that every dev team will eventually use such systems as part of their delivery process. He is confident about LaunchDarkly's potential to become one of the foundational tools in every developer's toolkit due to its superior capability compared to homegrown systems. He is excited to be part of LaunchDarkly's growing team and contribute towards making a significant impact on the world of software development.

The article discusses the integration of LaunchDarkly feature flags with AWS Lambda in serverless architecture applications. Serverless architecture offers faster development, lower costs, and less maintenance compared to traditional architectures. Feature flags allow teams to control feature releases without redeployment, which is particularly useful for stateless serverless applications. The article provides a step-by-step guide on how to integrate LaunchDarkly into an AWS Lambda application using the NodeJS runtime. It also explains how to store flag configurations in ElastiCache and process them within the application. Finally, it emphasizes that there are multiple ways to implement feature flags in serverless architecture applications depending on the app's structure, serverless provider, and implementation timeline.

Michael Gillett attended QCon London conference where he heard various talks ranging from microservice architectures to Chaos Engineering, focusing more on practices of creating software rather than showcasing frameworks and languages. A talk titled "Taking Back 'Software Engineering'" by Dave Farley stood out as it delved into the differences between being a Software Developer or an Engineer. Farley outlined three main phases of production we have gone through, emphasizing on Lean mass production and Just In Time (JIT) production which requires a good application of the Scientific Method to ensure high quality and efficient production lines. He argued that engineering and the application of the Scientific Method has led to some of humanity's greatest achievements, but there is a tendency in software development to shy away from the title of "Engineering". Farley suggested Hypothesis-Driven Development (HDD) as a way to truly move the industry to being one of engineers instead of developers. He also mentioned LaunchDarkly as an excellent tool for running experiments and testing hypotheses, allowing small changes to be made with minimal time and effort spent but useful results collected before major investment is made into developing a feature.

In a Test in Production Meetup, Steven Czerwinski from Scalyr discussed the importance of monitoring and observability while testing in production. He presented a case study on latency issues faced by one of their customers and shared key learnings realized after investigating the issue. Monitoring is crucial for testing in production as it helps shed light on hidden performance dark matter within complex distributed systems. Czerwinski emphasized following best practices in monitoring to avoid hindrances during investigations. He also highlighted the importance of consistency, analyzing based on what matters, and understanding averages versus distributions when designing monitoring systems.

In a recent Test in Production Meetup, New Relic Developer Advocate Clay Smith discussed instrumenting CI pipelines using the three pillar approach of monitoring - metrics, tracing, and logging. He observed a single run of a build with multiple steps kicked off by a code commit to explore what tracing looks like within a CI pipeline. Smith applied observability concepts to AWS CodePipeline while building Docker images and shared his findings on how these ideas can be used in understanding the work done by systems. The talk highlighted the importance of connecting production performance traces with build artifacts, such as git commit hashes or version numbers, for better visibility into complex pipelines and deploy strategies.

LaunchDarkly has been named #11 on the SaaS 1000 list for Q1 2018, marking it as one of the fastest-growing SaaS companies. The company raised $1 million in Series B funding in December 2017 and serves over 20 billion features per day to its customers. LaunchDarkly's platform allows various teams within a company to manage code in real time, enabling continuous software updates and improvements for better customer experience. CEO Edith Harbaugh was recognized as one of the top female entrepreneurs to watch in 2018.

At a recent LaunchDarkly Meetup, the focus was on monitoring and observability while testing in production. Priyanka Sharma from LightStep discussed how distributed tracing is an essential tool for testing in production due to the increasing complexity of software systems with CI/CD and microservices. Tracing helps teams understand their systems better and identify responses when testing in production. OpenTracing, a vendor-neutral open standard for distributed tracing, allows developers to instrument existing code without committing to an implementation from the start. It enables users to switch between different tracers and visualization systems easily.