July 2021 Summaries

Linedata revolutionized its software delivery lifecycle by adopting a self-service DevOps model and utilizing tools like Harness and Terraform, significantly reducing cloud expenses by 50% while enabling 150 to 200 daily deployments. The company shifted from a monolithic to a microservices architecture, implementing DevOps methodologies to enhance efficiency. Instead of relying on a centralized DevOps team, Linedata empowered autonomous teams to utilize and iterate on secure infrastructure templates provided by a global platform engineering team, under the guidance of DevOps Evangelist Andrey. This approach, combined with Continuous Infrastructure as Code, allowed for standardized and secure deployment processes, defined as "maturity institutionalized." To address cloud spend concerns, Linedata implemented FinOps, a model focusing on the efficient allocation of cloud resources, optimizing environments, and ensuring high ROI by balancing speed with financial oversight. Overall, Linedata automated its provisioning lifecycle, maintaining auditable and efficient operations through continuous feedback and integration with Git.

Harness introduced several updates in July 2021, including new server-side SDKs for Feature Flags in Python and .NET, expansion of Harness University with new courses, and future enhancements like robust dashboarding capabilities and Azure DevOps Repo support aimed at improving developer experience and operational efficiency. The updates also include new eBooks and articles focusing on the importance of developer experience, and upcoming events such as HarnessU Live sessions and industry summits. Harness is actively seeking early adopters for its new custom dashboard interface, which will offer more visualization options and integration with Azure DevOps for seamless synchronization. Additional features like enhanced event notifications, deployment freeze capabilities, and API support for approvals have been introduced to streamline DevOps processes, along with more granular access control to manage permissions effectively.

Single Digits successfully transitioned from a legacy deployment process to a modern microservice architecture using Harness, significantly improving their deployment efficiency and reducing costs. Initially, the legacy system required a complex and time-consuming deployment process, taking 6 hours every two weeks and involving off-hours sessions to minimize downtime risks. The adoption of Harness enabled Single Digits to reduce the deployment time to 20 minutes, saving $108,000 annually. Over 125 services have been onboarded, enhancing deployment velocity and minimizing downtime. The transition was spearheaded by Clint Nelissen, Director of Cloud Architecture, who sought a solution that would integrate with existing CI work without requiring a complete overhaul. Harness provided the necessary support to facilitate the migration, making software delivery a streamlined part of their modernization efforts.

Microservices architecture provides flexibility and scalability, but it demands careful management of DevOps processes such as CI/CD, monitoring, and logging to maximize its advantages. While the initial perception of microservices in 2016 was that of a buzzword, many organizations now benefit from transitioning away from monolithic applications, despite the challenges that microservices can introduce. This architectural style allows for independently deployable services that offer greater flexibility and cleaner overall architecture, although adapting existing frameworks or selecting new ones remains crucial for successful implementation. Various frameworks support microservices development, including Jakarta EE, Spring, Lagom, Dropwizard, and others like Vertx and Spotify Apollo, each with its own strengths and considerations. The choice of framework should align with the application's specific needs, and while microservices offer new possibilities, they require a balance of innovation and diligence in operational practices to ensure continued production support.

Measuring developer productivity involves a complex balance of metrics, frameworks, and tools to enhance efficiency while avoiding micromanagement. The SPACE framework, developed by Nicole Forsgren and researchers from GitHub and Microsoft, offers a holistic approach by focusing on satisfaction, performance, activity, communication, and efficiency, providing a nuanced view beyond simple output metrics. The article discusses the challenges and implications of measuring productivity, emphasizing that metrics should be aligned with both individual contributions and team dynamics. To improve productivity, it suggests integrating new tools within existing workflows, decoupling developer workstreams, and focusing on enhancing developer experience. Harness, a platform designed to boost developer productivity, simplifies CI/CD processes, reducing the burden on developers by allowing them to focus on writing code rather than managing pipelines, thereby minimizing context switching and potential burnout.

The text explores various Continuous Integration and Continuous Deployment (CI/CD) tools critical for DevOps, highlighting their unique features and capabilities. Harness is described as a versatile platform offering automation, deep integrations, and a focus on compliance, with support for multiple cloud services and robust governance features. GitLab is noted for its extensive features but criticized for requiring users to adopt the entire suite of GitLab tools. Codefresh is praised for modern features like pipeline as code and strong governance, though resource consumption and downtime are concerns. Azure DevOps provides customizable pipelines but requires extensive scripting, while Jenkins, despite its popularity, relies heavily on plugins and scripting. The text advises caution with certain tools like Travis CI, Jenkins, and Spinnaker due to limitations or complexity. Overall, the discussion underscores the importance of choosing the right CI/CD tools to streamline software delivery and enhance scalability.

The tutorial by Gabriel Cerioni focuses on deploying Vault Agent with Kubernetes Delegates using shared volumes to enhance secure and reliable token management in Kubernetes environments. This advanced guide builds upon a previous article about Vault Agent integration and Harness, offering steps for setup, configuration, and management of multiple Vault servers to ensure seamless integration and security. It emphasizes using ConfigMaps, PersistentVolumes, and Secrets, clarifying that the Vault Agent is a HashiCorp component, not part of Harness. The guide outlines a strategy for managing one Vault server per environment, detailing the use of Service Config Variables and Harness Environment Names to maintain atomicity and avoid single points of failure. The tutorial further explores creating a more professional Vault Agent deployment, handling security concerns by storing sensitive information in Google KMS Secrets Manager, and using Helm charts for Kubernetes Manifest configuration. It concludes with instructions for sharing tokens with Delegate deployments, ensuring both Vault Agent and Delegate reside in the same Kubernetes namespace, and verifying integration through logging and readiness probes.

Harness successfully transitioned from its homegrown feature flag solution, Local FeatureFlags, to Harness Feature Flags, effectively managing over 200 feature flags across six environments for more than 4,000 customers while maintaining simplicity for developers. This migration aimed to enhance the power of feature flagging without compromising ease of use, addressing limitations of the previous system which supported only basic boolean flags and required independent management for each environment. The transition followed a phased approach to mitigate risks, involving the use of a Java SDK for Admin APIs, running Harness Feature Flags in audit mode, and implementing an automated synchronization mechanism for new flags. The process included extensive testing and validation to ensure no performance impact, ultimately leading to a successful rollout in production. The transition has improved the feature flag management experience for various stakeholders, including developers, product managers, and QA teams, while maintaining a consistent and simple user experience.

Harness successfully transitioned from using an internal feature flag system to its commercial Feature Flags product, addressing challenges such as ambiguous flag naming and inadequate training. This shift not only eliminated the engineering burden of maintaining the old system but also improved functionality, security, and user experience. The migration was driven by a desire to "drink their own champagne," meaning they wanted to use their own products internally to enhance learning and feedback. Despite the difficulties, such as unclear support paths and insufficient communication, the company managed to overcome these obstacles and anticipates long-term benefits, including freeing up the Operations team to focus on more strategic tasks. The experience underscores the value of using one's own products to drive internal improvements and foster a more agile organization.

Gabriel Cerioni's guide provides detailed instructions for integrating Vault as a secrets manager with Harness, emphasizing the use of the Vault Agent to reduce manual efforts. The author shares personal experiences and challenges in setting up the integration, specifically addressing issues related to authentication methods and token life cycles. The process involves enabling the AppRole method, configuring the Vault Agent, and testing the integration with Harness. The guide includes steps to create roles and policies, retrieve necessary IDs, and configure a Vault Agent using a config.hcl file. Cerioni also highlights the importance of having the Harness Delegate access the sink file created by the Vault Agent. Although some features are currently behind a Feature Flag, broader availability is expected soon. The approach is portrayed as a secure and reliable method for managing secrets, with potential for further automation.

The blog post by Gabriel Cerioni provides a detailed tutorial on transforming a GitHub repository into an automated HTTP Helm Chart repository, which facilitates seamless integration with Harness for continuous delivery. By using a GitHub Action, the process automates the update of the repository index and simplifies Helm chart management, enhancing the efficiency and reliability of deployments. The tutorial outlines steps such as setting up a GitHub repository, creating a gh-pages branch for hosting Helm charts, and using the Chart Releaser Action to automate chart updates. The guide highlights the importance of maintaining a /charts path and updating an index.yaml file crucial for any HTTP repository. The author also shares insights on integrating the automated Helm Chart repository with Harness, offering a practical approach to managing Helm Repositories within software delivery workflows.

Role-Based Access Control (RBAC) is a security framework that enhances data protection by assigning user roles and permissions, ensuring selective access and separation of duties. RBAC is crucial for managing access in CI/CD pipelines and DevOps, as it offers flexibility and improved organizational security through regular role reviews and adjustments. Unlike access control lists (ACL) and attribute-based access control (ABAC), RBAC provides enterprise-level security by defining roles as collections of permissions, which can be adjusted based on organizational needs and structures. It supports businesses by protecting data and key processes, offering increased visibility across cloud tooling and IT systems, and facilitating the separation of duties to minimize the impact of cyber-attacks. Implementing RBAC requires a thorough understanding of organizational needs, role definitions, and regular reviews to maintain effective governance. The blog highlights how the Harness platform simplifies the application of RBAC in CI/CD environments, enabling authenticated user management through defined user groups and role assignments, thus enhancing software delivery practices.

Burst SMS, a global online messaging platform, significantly improved its software deployment process by adopting the Harness DevOps platform, which increased deployment frequency by eight times to once every two weeks and reduced deployment time to just five minutes. These changes transitioned the company from risky, infrequent "Big Bang" deployments every 3-4 months, which often resulted in production issues and approximately six days of annual downtime, to a more agile approach with minimal customer impact and downtime reduced to minutes. Harness provided Burst SMS with the ability to manage deployments efficiently across a hybrid cloud environment without requiring in-depth knowledge of specific cloud providers like AWS and GCP, fostering a culture of rapid iteration and innovation. This shift has allowed Burst SMS to embrace a "fail fast" mentality with reduced risk, demonstrating the value of DevOps practices within the organization.