June 2019 Summaries

The Harness Platform facilitates seamless integration of both legacy and modern technologies, such as Splunk and Prometheus, within continuous delivery pipelines, effectively addressing the challenges posed by the Pipeline Strangler effect. This effect, inspired by Martin Fowler's Strangler Pattern, describes how newer software architectures can overshadow and eventually phase out legacy systems. However, completely discarding older technologies is not always prudent, as organizations have significant investments in their existing pipelines. The text uses the example of Acme Corporation to illustrate how the introduction of new technologies like Prometheus, alongside established tools like Splunk, can create operational challenges. The Harness Platform offers a solution by simplifying the integration and management of various technologies within a pipeline, thus supporting innovation while maintaining critical legacy systems.

Harness and AppDynamics have integrated through the AppDynamics Integration Partner Program to enhance deployment velocity and performance verification within continuous-delivery (CD) pipelines, offering features such as automated rollbacks for performance issues and continuous monitoring with 24/7 Service Guard. This collaboration enables customers to leverage Harness for CD while using AppDynamics to verify application health, achieving accelerated deployments and performance verification in a single pipeline. The integration allows for mapping services to AppDynamics' Business Transactions, ensuring that any performance degradation triggers automatic actions such as rollbacks. Harness' 24/7 Service Guard continuously monitors services, leveraging machine learning on AppDynamics' performance metrics to maintain application health and support an interconnected IT ecosystem. This partnership is part of a broader effort to advance AIOps adoption and the realization of AppDynamics' Central Nervous System architecture, ultimately enhancing user experiences and accelerating time-to-market for enterprise customers.

Launching a new product feature effectively involves a structured five-phase strategy that has evolved significantly over the past decade, separating feature rollouts from code releases through the use of feature flags. These phases include dogfooding, debugging, Maximum Power Ramp (MPR), scalability, and learning, each serving distinct purposes such as internal feedback, risk mitigation, decision making, and understanding long-term impacts. Feature flags allow for controlled and experimental rollouts, testing features with varying user percentages to optimize performance and user experience while avoiding pitfalls like slow innovation or suboptimal outcomes. The strategy underscores the importance of a well-coordinated messaging and marketing plan to communicate the feature's value across multiple channels, leveraging social media to engage users and gather valuable feedback. By doing so, companies can ensure that technical and operational benchmarks are met while resonating with the target audience, ultimately maximizing market impact.