January 2023 Summaries

Harness has expanded its Chaos Engineering capabilities by integrating them with its Continuous Delivery platform to automate resilience testing and governance, helping teams identify system weaknesses and reduce downtime. This integration allows developers and SREs to test applications' reliability and resilience within software delivery pipelines, thereby improving velocity and minimizing unplanned downtime risks. As enterprises increasingly rely on cloud-native applications, the complexity of systems grows, making chaos engineering essential to prevent unforeseen downtime and technical debt. Harness Chaos Engineering, built on the open-source LitmusChaos project, enables automated chaos experiments with resilience scoring, ensuring systems are tested against various scenarios without impacting development speed. The Chaos Engineering platform supports serverless architecture resilience testing, with a significant expansion in its chaos fault library, including specific tests for AWS Lambda functions. This development is part of Harness's broader effort to create a unified software delivery platform, following its acquisition of ChaosNative Inc., which enables companies to build resilient and reliable software delivery pipelines rapidly.

Feature flags facilitate feature parity testing by allowing real-time comparisons between old and new systems, which aids in identifying hidden errors during system migrations and ensures a smooth transition with minimal risk. This testing method, also known as TAP compare testing, operates by mirroring traffic to both systems and comparing results, capturing discrepancies in actual traffic rather than relying on engineered test scenarios. To implement feature parity testing, it's crucial to have identical interfaces on both systems, which can be achieved by creating a matching interface or adding a façade. Feature management tools, such as Split by Harness, enable multiple traffic routing options like NEW, OLD, MIRROR, and PARITY, each serving different testing purposes. These feature flags are gradually rolled out to manage traffic and ensure infrastructure can handle the load before fully transitioning to the new system. This strategy is not only applicable to service migrations but also to data store transitions, where it helps verify successful data migration and system functionality. Despite the challenges, feature flags make parity testing more manageable, providing a critical security layer that can prevent overlooked features from impacting users, thereby enhancing the reliability of system upgrades.

Harness Chaos Engineering (CE) provides a comprehensive suite of tools designed to enhance system reliability by integrating chaos experiments into CI/CD pipelines, enabling unified experimentation across various platforms, and offering advanced observability features. These capabilities allow organizations to minimize unplanned downtime, improve software delivery, and proactively address system weaknesses by simulating real-world failures. Harness CE incorporates features such as steady state measurement, GameDay portals for team training, and robust experiment control methods, which help teams conduct safe and effective chaos testing. The platform supports integration with popular observability tools like Prometheus and Grafana, and offers enterprise-grade features such as audit trails, role-based access control, and extensive dashboards for clear communication and security. Built by the creators of the CNCF project LitmusChaos, Harness provides enterprise support to facilitate rapid adoption of chaos engineering practices, ensuring that businesses can scale their reliability efforts efficiently.

Harness has introduced Software Engineering Insights (SEI), a new module within its software delivery platform, designed to provide engineering leaders with actionable data to improve productivity, efficiency, and alignment. SEI, which integrates over 40 data sources including GitHub, Jira, and Azure, aims to enhance decision-making by offering insights into software delivery workflows, helping identify and remove bottlenecks, and improving execution, planning, and alignment with business objectives. The tool's integration with Propelo allows it to leverage established success metrics like DORA and the SPACE Framework, enabling leaders to manage performance effectively. SEI also features low-code automation to streamline processes and pre-built as well as customizable reports that facilitate data-driven decisions across diverse engineering environments. These capabilities allow companies to accelerate software delivery while maintaining quality and security, thus gaining a competitive advantage.

GitHub Actions is a CI/CD platform that automates software workflows, supporting multiple programming languages and integrating with various tools. It offers features like hosted runners, secrets management, and matrix builds, but also requires a significant amount of manual scripting and has limitations in continuous verification and metrics dashboards. GitHub Actions allows users to automate development workflows using YAML files, and it supports multiple operating systems through hosted runners, with a SaaS and on-premises version available for GitHub Enterprise users. While it is freely available for public repositories, private ones have a limited amount of free minutes for GitHub-hosted runners, with additional usage being chargeable. Despite its ease of use and flexibility in managing secrets, users face challenges with data passing between jobs, limited self-service capabilities, and a lack of native deployment tracking and metrics dashboards. Continuous verification functionality is limited, and although there are pre-written actions available, the platform's reliance on scripting can detract from valuable coding time. Additionally, managing canary deployments and secrets can be challenging due to limited control and integration with third-party tools.

Kubernetes labels are a crucial aspect of efficiently organizing and managing containerized applications, serving as metadata attached to objects like pods and deployments in the form of key-value pairs. These labels enable developers to enhance resource allocation, scheduling, and observability by allowing the selection and management of groups of objects based on specific characteristics. Best practices for using Kubernetes labels include employing descriptive and meaningful names, maintaining consistent naming conventions and values, avoiding excessive and overly generic labels, and leveraging label selectors to filter and manage resources. Advanced techniques such as label aliases, label policies, and label webhooks can further streamline label management, ensuring efficient deployment, scalability, and observability within Kubernetes clusters. Proper label management simplifies deployment and scaling, reinforcing the reliability and performance of applications while keeping the Kubernetes environment organized, efficient, and scalable.

Harness was recognized as the "Best End-to-End DevOps Tool/Service" in the 2022 DevOps Dozen awards, highlighting its leadership and innovation in the CI/CD space. This accolade, presented by Techstrong Group, a media company focused on the IT industry, emphasizes Harness's role as a comprehensive platform that significantly contributes to the DevOps community. Techstrong Group's annual awards celebrate the most innovative and influential individuals and companies in DevOps, cybersecurity, cloud-native, and digital transformation. The 2022 selection for the awards involved public nominations and a judging panel, with winners determined by public vote. Alan Shimel, Techstrong Group Founder, praised Harness for its continued progress and impact in the industry since its inception.

Securing Kubernetes workloads requires a comprehensive approach involving the use of trusted container images, role management, continuous monitoring, and integrating security tools like Kubescape, Datree, and Trivy into CI/CD pipelines. As organizations increasingly adopt Kubernetes to expedite software delivery, the distributed nature of this platform presents heightened cyberattack risks, necessitating robust security protocols to safeguard data and services. Kubernetes' security is rooted in the isolation of Linux containers, which limits attackers' access to the host operating system, but its open-source nature also makes it susceptible to additional vulnerabilities. Best practices for Kubernetes security include using trusted images, following the principle of least privilege, monitoring workloads for unauthorized access and data breaches, and leveraging CI/CD pipelines with integrated security tools to identify and mitigate vulnerabilities. As Kubernetes continues to gain traction, organizations are advised to balance speed with security by adopting DevOps principles and employing security scanners to protect their deployments against cyber threats.

End-to-end type safety in JavaScript is crucial for ensuring code correctness and reliability, particularly given the language's dynamically-typed nature. By utilizing tools like TypeScript, developers can introduce static type-checking to catch errors early in the development process, thus improving code maintainability and readability. Setting up a type-safe development environment involves using type-checking tools, adding type-checking to the build process, and enforcing runtime type safety with static type-checkers like TypeScript and runtime tools like prop-types. Best practices include defining interfaces for complex types, using type guards, writing unit tests, and leveraging static analysis tools to catch type errors before they reach production. Advanced strategies such as understanding type systems, using generics, managing dependencies, and building type-safe APIs further enhance type safety. Embracing a culture of type safety involves continuous education, code reviews, and automation to enforce standards, thereby fostering resilient and scalable software development. Additionally, platforms like Split enable controlled feature deployment and experimentation, enhancing development workflows and ensuring features are beneficial without introducing errors.

Harness has announced the general availability of the Harness Developer Hub (HDH), a comprehensive platform designed to enhance developer learning and expertise in software delivery by centralizing tutorials, documentation, and certifications across all Harness modules. Initially launched in beta, HDH now offers an expanded range of tutorials with hands-on examples, a migration of all documentation to this hub, and subscription options for release notes. The platform is structured into three main sections—Tutorials, Documentation, and Certifications—each aimed at facilitating a developer's journey from basic learning to achieving formal certification as a Harness expert. HDH encourages community contributions and plans to broaden its content based on user feedback, aiming to deepen feature documentation and expand its tutorial offerings to better support developers' needs in a dynamic software delivery environment.

Implementing Service Level Objectives (SLOs) and error budgets is crucial for balancing software delivery speed with reliability, thereby ensuring high service availability and customer satisfaction without striving for unattainable perfection. Site Reliability Engineering (SRE) teams play a pivotal role in maintaining application reliability through SLO management, which involves setting and monitoring target service levels to ensure applications function as expected. SLOs, along with Service Level Indicators (SLIs) and Service Level Agreements (SLAs), help organizations define and manage the reliability of their services, with SLAs being legally binding agreements with customers. Error budgets allow a margin for SLO violations, enabling teams to control release velocity and focus on service quality when necessary. Effective SLO management requires careful monitoring of metrics like latency, availability, throughput, and error rate, while avoiding over-alerting, to maintain a balance between innovation and reliability. The process is integral to delivering measurable and concrete reliability targets that facilitate both customer satisfaction and business competitiveness, acknowledging that perfection is neither possible nor desirable.