June 2022 Summaries

Harness has enhanced its support for AWS EKS Anywhere Bare Metal, allowing enterprises to deploy applications on bare metal servers in their data centers with the same efficiency and governance as cloud deployments, which is particularly beneficial for regulated sectors like public and financial services. This enhancement comes in light of AWS's recent expansion of its EKS Anywhere offering to include bare metal server management, enabling enterprises with spare physical node capacity to deploy Kubernetes clusters in their own data centers without additional licensing or complexity. Harness, a modern software delivery platform, uses Artificial Intelligence and Machine Learning to ensure applications are deployed successfully across both cloud and traditional infrastructures, providing rollback capabilities if necessary. The integration of Harness's platform with EKS Anywhere Bare Metal offers a synergistic solution that meets the stringent governance, security, and performance requirements of sectors that are unable to utilize cloud resources due to security or data sovereignty concerns.

Harness celebrated International Women in Engineering Day by showcasing the achievements of its female engineers and product leaders, emphasizing their contributions to technology and company culture, and encouraging more women to pursue careers in engineering. The company highlighted the importance of diversity, continuous learning, and community support in the tech industry, featuring insights from several of its standout female employees. Each woman shared personal stories about their career journeys, key skills for success, and the importance of women's representation in technology. They discussed the value of empathy, perseverance, and continuous learning, as well as the significance of overcoming self-doubt and building strong support networks. The celebration aimed to inspire more women to join and thrive in the tech industry, promoting a culture of inclusivity and growth within Harness and beyond.

Modern product development teams are increasingly embracing dynamic release models, leveraging feature data to deliver personalized and measurable experiences at scale. Techniques like feature flags, progressive delivery, and experimentation facilitate faster releases, enhance resilience, and enable precise decision-making, marking a shift from static release models. Emphasizing data-driven approaches, companies are redefining best practices in software delivery, prioritizing feature-level insights over application-level metrics. This shift enables the creation of numerous tailored versions of applications, fostering a culture of experimentation and continuous delivery. By utilizing feature management and context, teams can streamline rollouts, improve troubleshooting efficiency, and enhance customer support. The transition to a data-driven culture requires commitment and humility, allowing teams to quickly adapt to failures and make informed decisions. As development teams become pivotal knowledge sources for leadership, the focus on experimentation and data-driven decision-making positions them as frontrunners in the evolving landscape of software development and analytics. Through partnerships with platforms like Split, these teams are equipped to navigate and thrive in this new era of best practices, ensuring innovative and reliable product delivery.

Harness Service Reliability Management (SRM) is designed to help companies enhance application reliability through automating Service Level Objective (SLO) management, thereby preventing Service Level Agreement (SLA) violations and improving customer retention. The platform supports collaboration between development and reliability teams by providing a unified workspace and leveraging policy-as-code to set automated guardrails in the software delivery process. By addressing common challenges such as the manual management of SLOs and the slow identification of issues, SRM enables businesses to balance software reliability with innovation, reducing problem resolution time and avoiding penalties. Effective SLO management leads to increased deployment velocity, improved customer satisfaction, and protection against revenue loss, as demonstrated by Harness's customer Advanced, who benefited from quicker problem resolution and enhanced system reliability.

Harness celebrates Pride Month by hosting themed social events such as Drag Show Bingo and Pride Poetry Slam, and amplifying LGBTQIA+ voices to demonstrate its commitment to diversity and inclusivity within the tech community. The company encourages allyship by fostering a culture of inclusivity and providing opportunities to volunteer and donate to LGBTQIA+ communities worldwide. Employees like Geoff Taft, Luan Lam-Chen, and Chen Shterental share personal reflections on Pride, emphasizing the importance of authenticity, resilience, and allyship in fostering an inclusive environment. Pride Month serves as a reminder of the progress made and the ongoing challenges faced by the LGBTQIA+ community, with Harness inviting people to join their diverse team and live by their values of human connection and collective celebration.

Kubernetes is an essential tool for managing and scaling cloud-native applications, offering various service types like ClusterIP, NodePort, LoadBalancer, and Ingress to address specific networking needs. ClusterIP is used for internal communication within a cluster, NodePort exposes services on a node's IP at a static port, LoadBalancer provides a way to expose services to external networks, and Ingress acts as an intelligent router for directing external traffic to services based on defined rules. Each service type is tailored for different scenarios, making Kubernetes a versatile platform for efficiently managing containerized applications. As organizations increasingly adopt Kubernetes, understanding these service types is crucial for deploying and maintaining applications in a scalable and reliable manner.

The SPACE framework, developed by researchers at GitHub, the University of Victoria, and Microsoft, offers a holistic approach to measuring developer productivity by encompassing satisfaction, performance, activity, communication, and efficiency, thus addressing the limitations of traditional metrics like DORA. While DORA focuses on performance through metrics such as development velocity and quality, SPACE expands the scope to include factors that impact productivity at individual, team, and organizational levels. It emphasizes the importance of understanding developer well-being, the collaborative environment, and the efficiency of workflows, advocating for a comprehensive view that incorporates both quantitative and qualitative insights. By integrating these dimensions, SPACE aims to provide engineering leaders with nuanced data to enhance developer satisfaction and organizational success, moving beyond simplistic or isolated performance measures.

The text provides a detailed guide on deploying a Node.js application first on Minikube using local Docker images and then scaling it to larger Kubernetes deployments with Harness CD. It begins by outlining prerequisites, such as setting up Node.js, Docker Desktop, and Minikube, and explains the process of building local Docker images and deploying them on a Minikube cluster. The tutorial includes commands for starting Minikube, configuring Docker, building and deploying images, and accessing the application. For more extensive deployments, it introduces Harness CD, highlighting its ease of use for managing larger teams and more complex projects. The guide describes setting up a Harness CD project, configuring continuous delivery pipelines, and deploying to Kubernetes clusters, emphasizing the tool's user-friendly interface and seamless integration with existing workflows.

SPACE metrics provide a holistic framework for assessing developer productivity by considering performance, activity, satisfaction, and well-being, complementing the foundational DORA metrics which focus on velocity and quality. By integrating SPACE with DORA, organizations can enhance their insights and foster a more balanced approach to productivity and morale. The SPACE framework is an evolution of DORA, expanding its scope to include human and emotional factors that impact development teams. To effectively implement SPACE metrics, it is advisable to start with a selection of metrics from different dimensions to gain a comprehensive view of productivity, while also ensuring team involvement and transparency. The framework emphasizes the importance of viewing metrics as guidelines rather than absolute goals and encourages continuous improvement by adapting to organizational needs. Ultimately, SPACE provides engineering leaders with data-driven insights into developer satisfaction and productivity, enabling the development of sustainable and balanced practices.

Harness enhances the deployment of Azure Resource Manager (ARM) templates by introducing variables for templatization, which simplifies the management of microservices and offers rollback support, a feature absent in Azure. This approach allows for scalable infrastructure management by using a single parameter file for different microservices, reducing the complexity of maintaining multiple files. Harness variables, such as Service, Environment, Secret, and Workflow, are resolved before ARM deployment requests, enabling efficient template management. Additionally, Harness provides rollback support by saving the current state of the Resource Group before changes are made, allowing a return to the previous state if a deployment fails. This rollback mechanism is accomplished by deploying the existing template in COMPLETE mode, ensuring that only the specified resources remain.

Jim Sheldon explores various strategies for managing GitOps code, emphasizing the advantages and drawbacks of different repository structures, such as combining application and infrastructure code in a single repository versus using separate repositories for each environment. He argues that while a unified repository can simplify version control and reduce context switching for developers, it lacks privilege separation, which can be critical for some organizations. Conversely, having separate repositories for each environment offers better privilege separation and makes it easier to manage user access, albeit with a higher risk of configuration drift and increased complexity for developers. Sheldon suggests that the method of having one repository per environment is the most future-proof, providing flexibility to collapse repositories if separation becomes unnecessary. The blog also highlights the support offered by Harness' suite of products for various GitOps practices, encouraging readers to explore its features through a free trial or guided tour.

Extracting DORA metrics from Azure DevOps is complex and resource-intensive due to the limitations of Analytics Views and the extensive data manipulation required. Azure DevOps, a leading platform that integrates various phases of the development process, faces challenges in implementing DevOps analytics effectively. The extraction of DORA metrics, which quantify DevOps effectiveness through four key metrics—Deployment Frequency, Lead Time for Changes, Mean Time to Recover, and Change Failure Rate—requires significant effort, often akin to building a new product. The reliance on Power BI for data collection and the necessity for complex queries, coupled with the limitations in the types of data easily imported, make the process challenging. Furthermore, the unique structure of Azure DevOps complicates data aggregation and sharing, necessitating custom solutions for extracting meaningful metrics. The platform's inability to easily slice data by teams and the need for custom queries in Power BI add further complexity, suggesting that enterprises may need to consider commercial solutions for extracting DORA metrics effectively.

Chaos Engineering is a methodology that enhances system reliability by simulating failures to identify vulnerabilities, thereby improving performance and reducing downtime costs. Originating at Netflix in 2010, it has become an integral part of DevOps practices with tools like Chaos Monkey and LitmusChaos gaining widespread adoption. By intentionally injecting failure into systems, organizations can proactively address weaknesses before they result in costly outages, as evidenced by incidents like the Amazon blackout of 2018. This approach not only safeguards end-user experience but also bolsters team collaboration and incident response times. Companies such as Microsoft and Amazon have successfully integrated Chaos Engineering to maintain resilient distributed systems and microservices, with the practice offering opportunities for teams to learn and adapt to real-world failures. The ongoing evolution of Chaos Engineering tools continues to provide organizations with innovative ways to strengthen their infrastructure against unexpected disruptions.

Split is collaborating with industry leaders, including Dynatrace, in the OpenFeature initiative to develop a standardized API for feature flagging, aiming to reduce integration friction and enhance adoption across the tech ecosystem. This effort, submitted to the Cloud Native Computing Foundation for consideration as a sandbox program, seeks to establish a vendor-neutral standard that simplifies feature flag management, thereby facilitating progressive software delivery. Feature flagging, also known as feature toggling, allows for the decoupling of feature releases from deployments, enabling runtime control over functionality without new code deployments. OpenFeature aims to standardize this process by providing a common API and SDK, fostering a more seamless integration across diverse tech stacks and encouraging broader usage by lowering entry barriers. Split supports this initiative by offering advanced targeting for controlled rollouts, flag-aware monitoring, and feature-level experimentation, emphasizing the importance of data-driven feature management. The initiative is seen as a pivotal step towards making feature flagging a mainstream best practice in software development, promoting good software practices through a collaborative, vendor-neutral approach.