February 2023 Summaries

LaunchDarkly, a feature management platform, has released new updates to help businesses deliver software faster with less risk while maximizing business value. The release includes custom contexts for more targeted feature releases, an approvals dashboard for better change management tracking, and updated notification settings for streamlined communication. Additionally, the platform now supports Microsoft Teams integration, RUM integrations for Datadog and AWS users, SCIM support for Azure Active Directory, team flag maintainers, a dark theme option, and has achieved FedRAMP authorization for government use.

LaunchDarkly introduces Custom Contexts, a new core capability that allows users to define, experiment, and target based on custom contexts relevant to their businesses. This feature enables software deliveries with unlimited flexibility by directly reflecting the data model, business language, and release process of the user's organization. With Custom Contexts, LaunchDarkly has moved from one-dimensional, user targeting to multidimensional targeting called "multi-contexts." The new feature allows users to deliver features to various entities such as accounts, projects, sessions, etc., all at once, increasing power and control while reducing risk and technical debt. This capability is essential for complex software deliveries and has become a foundational requirement for feature management.

AWS held its annual Re:Invent conference in late November 2022 where John Kodumal, CTO and co-founder of LaunchDarkly, delivered a session titled "Building and Operating at Scale with Feature Management." The session emphasized that feature management is not just about showing or hiding features from end users; it's also a way to evolve software safely. Kodumal discussed how feature flags can be used for application modernization, compliance, and cost performance optimization. He shared real-life examples of using LaunchDarkly for these purposes, including migrating data stores, ensuring HIPAA compliance, and staging the rollout of new technologies while maintaining performance and uptime. The session provided valuable insights into how feature management can be leveraged to improve software development processes.

LaunchDarkly, a feature management company founded in 2014 by Edith Harbaugh and John Kodumal, has grown to $100 million ARR and serves over 4,000 customers. The company is now transitioning into its next phase of growth with Dan Rogers taking on the role of CEO. Previously, Rogers served as president at Rubrik, where he helped achieve $500M ARR in under nine years, and as CMO at ServiceNow, contributing to their rapid growth journey. Harbaugh will continue to be involved with LaunchDarkly by stepping into a new role as Executive Chair of the board, focusing on customer relationships and promoting the company globally.

Feature flags are configuration settings that allow developers to enable or disable specific features without modifying code. They provide a fast, easy way to control feature releases, testing, and deployment. Feature flags can be used in various scenarios such as A/B testing, permission management, and microservice architecture. By using feature flags, developers can prevent issues caused by rapid versioning and maintain granular control over their applications. However, proper naming, organization, and maintenance of feature flags are essential to avoid confusion and ensure code coverage during testing.

At AWS Re:Invent conference 2022, Andrew Krug from Datadog and Alex Hardman from LaunchDarkly led a session on observability, security, and feature management. They discussed how threats have evolved and new detection methods, emphasizing the role of feature flags in detecting and responding to attacks. The talk covered creating a robust security culture within an organization, incident response process acceleration, leveraging lightweight threat modeling, and adversary engagement strategies. They also demonstrated how Datadog and LaunchDarkly integrations can be used for these purposes.

Feature flags are a crucial tool for managing complex codebases and ensuring smooth deployments while maintaining application availability. They allow developers to release features in a disabled state, enabling them only when the feature flag is turned on. This decouples deployments from functionality releases, giving software teams more control over their applications. Site reliability engineering (SRE) teams can also benefit from feature flags by using them for testing performance impact, temporarily disabling features to maintain system stability, performing gradual rollouts of new features, and observing metrics changes. Feature flags empower SRE teams to work independently and effectively in maintaining the performance and stability of applications without interfering with developers' tasks.

The blog post series "Galaxy Brain" discusses progressive delivery in the context of software development, emphasizing the importance of quality over speed. Rob Jahn, a DevOps Evangelist at Dynatrace, highlights that manual processes often hinder progress and suggests using metrics provided by Dynatrace during the development lifecycle to ensure quality throughout the process. Automating release validation with service level indicators (SLIs) and service level objectives (SLOs) is also recommended for a quick feedback loop. The post concludes with an overview of how Dynatrace can be used to compare feature settings, analyze error rates and response times, and integrate with other tools like LaunchDarkly.

A software release is a set of updates or new features for a product or service. Major releases introduce significant changes and may impact user experience, while minor releases fix bugs or improve performance. Emergency releases address critical issues outside the regular release cycle. The software release lifecycle includes stages such as pre-alpha, alpha, beta, release candidate, and final release. Release management has evolved from project management to driving business value through identifying bottlenecks and analyzing user interactions. Deployment strategies like canary deployments and feature flags allow for decoupling deployment from releases, enabling better testing and rollback capabilities.

HashiCorp's Terraform is a popular tool for infrastructure provisioning, but its capabilities extend beyond just that. Users can choose from hundreds of providers to manage various resources, including ordering pizzas or building Spotify playlists. The LaunchDarkly Terraform provider allows users to utilize it for their feature management needs. To ensure security when using the LaunchDarkly Terraform provider, sensitive variables should be marked as such and state files containing sensitive information should be stored in a secured remote location like Amazon S3 or Terraform Cloud. For added security, HashiCorp's Vault can be used to store and retrieve API access tokens. It is recommended to break apart configurations into separate files for easier management and adjustments. Descriptive tags and descriptions should be added to resources for better understanding of changes made in the UI. Terraform can also be used beyond provisioning, allowing users to make additional changes to their LaunchDarkly projects. Lastly, HashiCorp's Terraform Cloud offers collaboration benefits and integration with VCS like GitHub or GitLab, enabling automatic planning of proposed changes as code is being worked on.

Mass layoffs in tech companies have led to concerns about loss of institutional knowledge and increased workloads for remaining employees. Laying off a significant portion of engineering staff can result in longer incident remediation times, fewer bugs being fixed, and an increase in technical debt. The remaining engineers may need to take on tasks outside their current skillset, leading to a challenging work environment. Additionally, the loss of specialized knowledge can affect ongoing projects or future plans, making it difficult for companies to maintain their competitive edge.