February 2021 Summaries

Harness is presented as a robust alternative to Octopus Deploy for continuous delivery, offering features such as machine learning-based deployment verification, deep integrations, and compliance tools that enhance deployment reliability and reduce setup time. Octopus Deploy, a deployment automation tool founded in 2011, is praised for its reliable deployments and integration with TeamCity but criticized for limited third-party integrations and high setup efforts. Competitors such as Azure DevOps, GitLab, and Jenkins are evaluated, highlighting Azure's flexibility but scripting demands, GitLab's comprehensive DevOps platform with a learning curve, and Jenkins' reliance on plugins and outdated features. Harness is emphasized for its comprehensive suite, including Continuous Integration, Continuous Delivery, and Cloud Cost Management, supporting major cloud platforms and offering innovative compliance and governance capabilities, while also being involved in the open-source community through the acquisition of Drone.

Harness celebrated several achievements in February, including the integration of Google Secrets Manager with its continuous deployment (CD) platform, the success of its ShipTalk podcast, and plans for future advancements like Azure support and Community Edition What If Analysis. The company expressed excitement for March, highlighting upcoming SkilUp Day events and the ongoing CI Enterprise Beta, which has received positive feedback since its launch in January. February also saw the release of three new podcast episodes, covering topics such as DevOps transformation, open source community development, and overcoming imposter syndrome. The company has been featured in the press for insights into successful DevOps leadership and managing Kubernetes complexity. Looking ahead, Harness is preparing for significant updates in the spring, with a focus on expanding Azure support and releasing the Harness Community Edition 2.

Implementing DORA metrics, derived from six years of research by the DevOps Research and Assessments team, allows organizations to measure and enhance software delivery performance by tracking deployment frequency, lead time for changes, change failure rate, and time to restore service. These metrics provide insights into the velocity and quality of software engineering outputs, although gathering data for them can be complex due to the variety of tools used across teams. Platforms like Harness Software Engineering Insights offer out-of-the-box dashboards and predictive analytics to facilitate decision-making and accommodate diverse DevOps toolchains. These platforms not only standardize software metrics but also offer flexibility for organizations to extend metrics and insights to various audiences, from executives to engineering teams, enhancing their ability to visualize and act upon these analytics.

GitOps has become an influential advancement in the realm of DevOps by employing Git as a single source of truth to streamline continuous deployment for cloud-native applications and programmable infrastructure. By integrating version control with automation, GitOps allows for seamless deployment into environments like OpenShift or Kubernetes, enhancing collaboration between development, QA, and operations teams. It builds on the simplicity of Git, making it accessible to engineers familiar with version control processes, while offering significant benefits such as ease of rollback and deployment history tracking. However, challenges remain, particularly in managing multiple repositories and secure storage of secrets, which necessitate careful handling and additional tools for centralized secret management. Despite these challenges, GitOps is seen as the maturation of DevOps practices, promising more efficient and repeatable processes across the software development landscape.

The walkthrough details the process of creating a type-safe REST API using Node.js and TypeScript, integrating feature flags with Split to facilitate controlled feature delivery and A/B testing. It emphasizes the advantages of using TypeScript, a type-safe superset of JavaScript, to enhance the development experience by providing type safety and facilitating the creation of scalable applications. The tutorial guides users through setting up a TypeScript development environment, building an Express server, creating a feature flag with Split for a GET /timezones endpoint, and utilizing middleware for dynamic feature management. It also covers the use of declaration files, the benefits of using feature flags in real-world applications, and how to manage code branching with feature flags for a consistent user experience. The example demonstrates how to apply these techniques in a Node.js environment to improve software development practices and enhance application reliability and user experience.

The blog post offers a comprehensive overview of artificial intelligence (AI) and machine learning (ML), differentiating between the two and exploring their various types, including supervised, unsupervised, and reinforcement learning. It emphasizes the importance of understanding AI and ML due to their widespread and often misrepresented use in marketing. The author highlights practical use cases across sectors like transportation, medicine, and software development, asserting that AI and ML have a reciprocal relationship with DevOps, as illustrated by MLOps and tools like Kubeflow. The post also provides a practical example of data preparation and model training using Python libraries such as pandas and NumPy, focusing on a diabetes dataset to demonstrate the process of cleaning data, training a model, and evaluating its performance. The content is based on a presentation given at a DevOps meetup, aiming to encourage further exploration of AI and ML.

Logging engine performance varies significantly across various tasks, with Log4J2 demonstrating remarkable speed in handling exceptions, underscoring the necessity of selecting the appropriate logging tool based on specific application requirements to optimize performance and resource usage. Logging is crucial for server-side applications, acting as the primary method for outputting live state in a persistent and readable manner. The study examined multiple prominent logging engines, including Log4J, Log4J2, Logback, SLF4J, and Java Util Logging, comparing their performance in standard logging activities such as logging string constants, POJOs, and throwable objects. These tests revealed that Log4J2 excelled particularly in logging exceptions, while the performance varied across other tasks. The experiments, conducted on an Amazon m1.large EC2 instance, highlighted how configuration changes, such as adjusting Log4J2's layout, can significantly impact performance. The results emphasize the importance of understanding the costs associated with logging context data and choosing the right engine to meet specific logging needs effectively.

Bart Farrell's journey from non-technical fields such as theology and talent management into the tech industry, particularly the Data on Kubernetes Community, underscores the significance of embracing change and building confidence through incremental successes. Despite not having a traditional tech background, Bart's story of navigating the complexities of technology and imposter syndrome is a testament to the power of community support and the human element in professional growth. The Data on Kubernetes Community, which Bart leads, provides a platform for individuals with varied expertise to share experiences, discuss challenges, and explore the possibilities of running stateful workloads on Kubernetes. The community thrives on inclusivity and collaboration, encouraging participation from people of diverse backgrounds and skill levels. Bart emphasizes the importance of not needing to have all the answers or being an expert to contribute, as the community can help individuals complete their stories, reflecting a supportive environment for personal and professional development.

Continuous Integration (CI), Continuous Delivery (CD), and Continuous Deployment (CD) are interrelated processes that automate software delivery, each with distinct goals and methodologies. CI focuses on automating the building and integration of code changes, ensuring that they are consistent, repeatable, and available, while CD emphasizes automating the deployment and release processes to safely deliver software changes to production. Continuous Deployment automates the actual deployment of code to production environments once it passes quality checks, minimizing human intervention. Together, these practices aim to streamline the software delivery pipeline, allowing for faster, more reliable updates and features in production, with automation being key to reducing manual tasks and enhancing efficiency. The processes are supported by modern tools and platforms like Kubernetes and the Harness Platform, which facilitate scalable, agile, and safe software deployment across various environments.

Platform-as-a-Service (PaaS) simplifies software development by offering ready-to-use application infrastructure, allowing developers to focus on building applications without worrying about the underlying environment. Originally popularized by Heroku, PaaS has evolved to include major offerings such as OpenShift and Cloud Foundry, providing a more structured and opinionated approach to software deployment. The integration of PaaS with tools like Harness enhances Continuous Delivery by automating deployment and rollback processes, making it easier to manage applications across hybrid and multi-cloud environments. However, the rise of Kubernetes as a dominant orchestration platform has challenged traditional PaaS offerings by providing a more flexible and scalable solution for cloud-native applications. PaaS vendors have begun to incorporate Kubernetes to stay competitive, with companies like Red Hat's OpenShift embracing it as an orchestrator. Despite the competition, PaaS still offers unique advantages in terms of portability and developer experience, with Harness supporting seamless integration to maximize technology investments.

Implementing continuous delivery practices allows businesses to deploy software more efficiently and reliably, which reduces downtime and accelerates time to market. By using tools like Harness, organizations can achieve seamless deployments across various environments, such as AWS Lambda, thereby enhancing their operational agility.

In an episode of ShipTalk, Ravi Lachhman speaks with Tracy Miranda, the Executive Director of the Continuous Delivery Foundation (CDF), about building sustainable and inclusive open-source communities. Tracy emphasizes the importance of having a clear vision, being deliberately inclusive, and participating in programs like Google Summer of Code to attract diverse contributors. She shares her experiences from her time at CloudBees and the Eclipse Foundation, highlighting the challenges of fostering open communities and the significance of transparency and shared goals in overcoming them. Tracy also discusses the value of open-source contributions in advancing individual careers by creating a transparent portfolio and connecting with a supportive network of professionals. She advises those starting new open-source projects to focus on defining their goals, seeking out like-minded individuals, and ensuring inclusivity from the outset. Additionally, Tracy encourages engineers to leverage open-source opportunities for professional growth and to engage with the community through events like cdCon, the CDF's annual conference.

Harness CD's Deployment Freeze feature provides managers with the ability to schedule blackout windows, preventing code deployments during critical business periods and thereby reducing downtime risks that could impact revenue. These scheduled deployment freezes, or blackout windows, ensure that high-stakes business operations, such as retail events like Black Friday or financial firms' peak traffic times, are not disrupted by new code deployments. This feature allows companies to avoid ad-hoc scheduling mishaps and offers a structured approach to managing deployment pipelines. Before this feature, deployment freezes were managed manually, often leading to communication errors. With Harness CD, only admins can set up these freezes via the governance tab under security, ensuring that deployments are paused during specified times, with clear notifications to those attempting to deploy during these periods.

Cloud cost management tools are crucial for optimizing cloud expenses by providing visibility into spending, identifying unused resources, and offering cost-saving recommendations. These tools are essential for finance, engineering, and FinOps teams across various cloud environments, such as AWS, GCP, Azure, multi-cloud, or hybrid cloud settings. They help address issues like unused or overprovisioned resources, 'zombie' resources, and offer solutions ranging from reserved capacity purchases to cultural changes in cost management. The article discusses several tools, including Harness Cloud Cost Management, AWS Cost Explorer, GCP Billing, Azure Cost Management + Billing, Apptio Cloudability, CloudHealth by VMware, Spot by NetApp, and Kubecost, highlighting their pros and cons. While some tools excel in automation and multi-cloud visibility, others require good tagging practices for granular insights. Ultimately, selecting the right tool depends on the company's specific needs and scale, with Harness Cloud Cost Management positioned as a modern solution that integrates well with broader software delivery platforms.

Instrumentation is a critical component of monitoring application performance within the DevOps framework, enabling teams to detect errors, ensure predictability in release cycles, and enhance user experience through actionable insights. It involves the use of Application Performance Management (APM) tools, daemons, and log management to provide comprehensive views of an application's state and health. Key to successful instrumentation is understanding what to measure and how it aligns with both technical and business processes, as outlined by James Fryman at Monitorama 2016. Fryman emphasizes the importance of following a structured approach, akin to a DevOps version of Maslow's hierarchy of needs, starting with a culture of learning and experimentation. Despite potential pushback, particularly due to past unproductive attempts, iteration and developer collaboration are crucial for refining instrumentation practices. Additionally, the text highlights the use of Java agents and daemons like StatsD for data collection, and the role of log management tools in streamlining the analysis of extensive log data to improve application reliability.

Value stream mapping (VSM) is a lean manufacturing technique that has been adapted for software development to optimize the flow of materials and information within an organization's processes. It originated from the manufacturing industry, particularly with Toyota, and has since been integrated into DevOps practices to improve software delivery. VSM provides a visual map of the current state of a process, helping teams identify bottlenecks and areas for improvement by examining the flow of information, materials, and time. It involves key stakeholders to ensure a comprehensive understanding of the workflow and employs symbols to represent different components of the process. The technique also incorporates lean principles like muri (overburden), mura (unevenness), and muda (waste) to streamline operations. Challenges such as ensuring the right participants are involved and distinguishing between current and future state maps are crucial for the success of VSM exercises. When executed effectively, VSM workshops can drive change by highlighting outdated tools and processes, ultimately accelerating software delivery processes.

Maintaining a high hygiene score, ideally 70 or above, in engineering processes is crucial for ensuring clean and trustworthy data, which in turn supports informed decision-making and fosters more productive and organized teams. The Harness Software Engineering Insights module aids in achieving this by providing out-of-the-box hygiene factors and reports, enabling teams to quickly assess and enhance their processes. It draws an analogy between personal hygiene habits, like frequent hand washing during COVID, and engineering hygiene, emphasizing the importance of following process guidelines to ensure data cleanliness. Hygiene factors include practices such as adding JIRA ticket information in Pull Requests and ensuring tickets are properly groomed. The module allows for easy customization of hygiene factors and utilizes tools like JIRA, GitLab, and PagerDuty to measure hygiene scores. A high hygiene score correlates with better team performance, while poor hygiene can lead to issues like retention problems and poor execution. Organizations are encouraged to assess their hygiene score and consider using Harness to improve engineering success.

Advancing from middle management to senior leadership in the technology sector involves embodying traits such as vulnerability, transparency, data-driven decision-making, decisiveness, empathy, and effective personnel management. The author, who rose through the ranks from Cisco to ThoughtSpot, emphasizes the importance of self-reflection and openness about challenges, which builds trust and credibility with both superiors and teams. Transparency in sharing the rationale behind decisions fosters trust and alignment toward common goals. Data is crucial for making objective decisions, removing biases, and improving processes, which complements instinct and experience. Decisiveness is key, as leaders must make timely decisions even with limited information. Empathy is vital, especially in remote work settings, to respect and support team members, while timely letting go of underperforming employees is necessary to maintain team dynamics and productivity. Overall, successful engineering leadership combines soft skills with strategic, data-informed decision-making to build thriving organizations.

Vuclip, a global media company focusing on emerging markets, transitioned from monolithic on-premises applications to cloud-native microservices using Google Cloud Platform, significantly optimizing their deployment processes with the help of Harness. The adoption of Harness reduced deployment time from 24 hours to 4 hours and automated rollbacks from 2 hours to just 15 minutes, leading to an expected annual saving of $200,000 and accelerating Vuclip's cloud migration by four months. Before Harness, Vuclip's deployment process was manual and time-consuming, involving centralized teams and extensive engineering hours. Harness's integration capabilities with Kubernetes, New Relic, and ELK, as well as its user-friendly interface, allowed Vuclip to automate deployments and empower individual development teams to manage their own processes. This integration not only reduced the workload but also enhanced deployment verification and rollback efficiency, preventing unplanned downtime and avoiding the need for additional engineering investments in building custom CD capabilities.

Linedata, a global technology solutions provider for the asset management and credit industries, significantly improved its deployment process by implementing Harness Continuous Delivery, achieving an 80% reduction in deployment times and an 88% faster verification/testing period, while saving $500,000 in DevOps costs. The company transformed its process by adopting a microservices architecture based on .NET core running in the public cloud, moving away from manual deployment scripts that were time-consuming and inefficient. This transformation included a self-service Continuous Delivery initiative, enabling developers to trigger automated deployment pipelines from TeamCity, which streamlined the promotion, testing, and verification of new builds across all environments. The integration of Harness has empowered teams across multiple geolocations, enhancing collaboration, reducing time-to-market, and ultimately improving the product experience for customers.

Lavasoft, a subsidiary of Avanquest Software, significantly improved its software delivery process by transitioning from complex Jenkins pipelines to Drone CI and Harness CD, leading to a 97% reduction in upgrade times and a 94% decrease in verification time. This change facilitated faster Kubernetes migrations and enhanced deployment visibility, allowing product managers better access to deployment schedules and metrics. Previously, Lavasoft faced challenges with its diverse application languages and on-premise hosting, making it difficult to apply DevOps principles and resulting in frequent manual interventions for deployment issues. By adopting Drone CI, which simplified the integration process with its containerized plugins, and Harness CD, which offered compatibility with Kubernetes, Lavasoft reduced the time spent on troubleshooting and improved communication between product managers and developers. These tools enabled seamless onboarding of applications, reduced migration times, and integrated continuous verification tools like Datadog and ELK, enhancing the overall efficiency of Lavasoft's software delivery system.

Relativity successfully transformed its deployment pipelines by implementing Harness CD, resulting in a 30-fold increase in deployment velocity, a reduction in deployment time from 8 hours to minutes, and an 80% decrease in the change failure rate, all while saving $150,000 by avoiding the need for dedicated engineering resources. Led by Corey Wagehoft, the initiative involved migrating from legacy Windows services to container-based microservices in Microsoft Azure, which required overcoming security challenges and the limitations of previous tools like Spinnaker. Harness CD was chosen due to its compatibility with both Azure microservices and native Windows services, integration with Azure Key Vault for secrets management, and secure communication via its delegate architecture. Within three months, Relativity enhanced its deployment process with Harness, integrating with existing tools like Jenkins, Terraform, and Splunk, and enabling consistent and repeatable deployments through pipeline templates. This transformation not only improved operational efficiency and compliance with security standards but also empowered teams to make data-driven decisions regarding pipeline flow and deployment metrics.

By adopting the Harness platform for continuous delivery, Advanced, the UK's third-largest provider of business software and services, achieved a 10X return on investment in just two months, significantly enhancing its deployment efficiency. The company, which supports over 16,000 customers with 2,200 employees, transitioned from traditional on-premise to AWS cloud solutions, initiating an internal CI/CD initiative to empower its 700 software engineers across 100 product teams. Previously, deployment processes were cumbersome, requiring extensive manual intervention and often taking months to build and hours to execute. Harness provided a self-service platform that streamlined deployments, reducing average deployment time from two days to two hours and onboarding time from two days to 90 minutes, while also automating tasks such as Jenkins upgrades. This resulted in a 50-60% reduction in DevOps and engineering time spent on deployments, allowing teams to templatize and reuse deployment pipelines, significantly cutting time-to-market and enhancing overall productivity.

Jobvite, a leader in recruiting software, transformed its complex deployment process into an efficient Continuous Delivery (CD) system using Harness, reducing deployment times by 90% and achieving an estimated annual saving of $500,000. This transition allowed developers to deploy their code directly into production, enhancing the quality of service and customer experience. Initially, the CD process was semi-manual and maintained through custom scripts, which posed challenges as the core application and technologies evolved. The implementation of Harness facilitated canary deployments and integration with existing DevOps tools, improving stability and reducing deployment time from 27 minutes to just 2 minutes. This change empowered developers to manage their own code while the release team focused on automation, making Harness central to Jobvite's operations and presenting a clear return on investment compared to building a similar system in-house.

Marketron, a leading provider of revenue management solutions for media outlets, transformed its software delivery process by adopting Harness, significantly increasing its deployment frequency and reducing deployment times. Previously, Marketron struggled with infrequent bi-annual deployments, relying on Jenkins and manual processes, which resulted in a labor-intensive 4-hour downtime window and a preference to avoid frequent updates due to potential revenue-impacting issues. Under the leadership of Mike Jackson, the new SVP of Software Engineering, the company initially doubled its deployment frequency to quarterly releases, but this proved insufficient for their goals, particularly in light of a planned migration to AWS and Kubernetes. Faced with the choice of hiring more staff or improving their tools, Marketron chose to implement Harness, which enabled them to increase production deployments to a monthly schedule and slash deployment time to just one hour. This change also facilitated 1200 non-production deployments monthly, enhancing testing and service quality, and marking a significant modernization of their software delivery process.

Casting Networks, a leading entertainment technology company, significantly accelerated its transition to Kubernetes microservices by adopting Harness Continuous Delivery, a move that reduced microservice onboarding time by 97% and deployment time by 95%, while also decreasing the change failure rate by 90%. Initially facing challenges with their legacy Jenkins deployment pipelines, which were complex and time-consuming, the Director of DevOps, Jeremy Malara, recognized the need for a more efficient solution to enhance developer velocity and application stability. Harness offered a standardized, self-service approach to deployment, allowing developers to easily manage and deploy their microservices, thus eliminating bottlenecks and reducing reliance on a small team of Site Reliability Engineers (SREs). The implementation of Harness not only streamlined the migration process but also embedded Continuous Verification into pipelines, facilitating automated QA and canary deployments. This strategic shift enabled Casting Networks to advance their Kubernetes migration timeline by four months, ultimately supporting their business objectives more effectively and efficiently.

Tilting Point, a prominent free-to-play games publisher, transitioned from Amazon Elastic Beanstalk and Jenkins to Kubernetes with the assistance of Harness, resulting in a threefold increase in deployment frequency and significant cost savings by eliminating the need for a dedicated DevOps engineer. Previously, Tilting Point's lead software engineer, Evan Thomas, was the sole person managing the CI/CD pipelines, which were prone to rolling deployment risks and false monitoring alerts, consuming a substantial portion of his time. The move to Kubernetes, facilitated by Harness, empowered Thomas to modernize the company's applications and achieve deployment efficiencies similar to those seen in larger tech companies with extensive DevOps teams. This strategic shift has enabled Tilting Point to enhance their software delivery processes without expanding their team, supported by the ongoing partnership with Harness's support team.

Beachbody significantly enhanced its cloud infrastructure and deployment efficiency by leveraging Harness Continuous Delivery, which notably accelerated its AWS cloud migration from one year to three months and doubled production deployment velocity. The adoption of Harness reduced QA testing time by 92% and deployment effort by 90%, while also cutting application onboarding time from three days to just two hours, ultimately deferring $600,000 in DevOps costs. Previously reliant on Jenkins Pipelines, Beachbody faced challenges with manual, costly, and time-consuming processes, which have now been streamlined with Harness's automation capabilities, including integration with tools like GitHub, Terraform, and Dynatrace for seamless deployment and monitoring. The company's transition to cloud-native services with Kubernetes and Lambda has facilitated automated, self-service deployment with a robust security and compliance framework, supported by Harness's role-based access control and audit trail features, enabling rapid pinpointing of deployment impacts.