December 2019 Summaries

Since joining GitLab in 2018, the author has navigated the complexities and opportunities of working in an all-remote organization, highlighting both challenges and solutions in hopes of aiding others new to remote work. Despite missing the energy of physical human interaction, the author found that tools like Slack and Zoom effectively simulate office camaraderie, even offering deeper connections through virtual visits to colleagues' workspaces. Initially struggling with productivity due to a lack of social validation, the author embraced the autonomy GitLab offers, defining clear daily objectives to foster a sense of freedom and work-life balance. While overwhelmed by meeting-heavy schedules, the author utilized GitLab's culture of recording and optional meetings to manage their time more effectively. A significant adjustment was the emphasis on documentation, which encouraged the author to seek solutions independently and contribute to the company handbook, enhancing self-reliance. Overcoming discomfort with video conferencing, the author learned to focus on the work rather than the background, fostering stronger relationships. As remote work becomes more common, the author emphasizes the importance of sharing experiences and solutions within the remote community.

A Fortune 50 company sought guidance from GitLab's CEO, Sid Sijbrandij, on enhancing communication, product management, and project execution, leading to the sharing of GitLab's best practices. These include utilizing directly responsible individuals (DRIs) to streamline decision-making without waiting for consensus, clearly defining product and engineering responsibilities to avoid accountability issues, and fostering collaboration through InnerSourcing by making internal code and issue trackers open to all. GitLab emphasizes the importance of documentation to prevent repeated discussions and maintain a clear record of decisions, while iteration is encouraged to allow for quick feedback and reduce coordination complexities. The "jobs to be done" framework is employed to align product development with customer needs, ensuring prioritized improvements. These practices, central to GitLab's success as an all-remote company, are adaptable for any organization aiming to improve workplace communication or transition to a remote setup.

GitLab's "public by default" approach has led to the development of Token-Hunter, a tool designed to identify sensitive data unintentionally shared in public GitLab assets such as issues, snippets, and merge requests. Recognizing the common pitfalls of exposing API tokens and other sensitive information in shared environments, GitLab's Red Team created Token-Hunter to complement existing tools like gitrob, TruffleHog, and gitleaks. The tool is engineered to efficiently search for and identify exposed data through the use of request retries and dynamic page-size reduction algorithms, successfully analyzing nearly 1.3 million GitLab assets. Token-Hunter allows users to tune regular expressions based on their specific environments to improve search accuracy. GitLab encourages community contributions to enhance the tool's features, such as better output formatting, real-time reporting, and data persistence, aiming to support the broader open-source community in defending against potential security threats.

The blog post explores potential improvements to GitLab's merge request interface, focusing on enhancing code review efficiency and collaboration. It highlights recent updates like Sourcegraph integration and multiple assignees, aiming to streamline user navigation and performance. The post proposes ideas such as real-time collaboration features, similar to Google Docs, to enable simultaneous work on merge requests and improve communication. It emphasizes the importance of integrating external discussions directly into the merge request platform to maintain a single source of truth and reduce tool redundancy. The article invites feedback on these concepts to address privacy concerns and optimize the balance between real-time collaboration and individual workflow, aiming to create a more efficient and collaborative code review environment.

GitLab's engineering team conducts monthly retrospectives to foster continuous improvement and prevent technical debt from outpacing code development. These retrospectives are pivotal in enhancing transparency, trust, collaboration, and communication within the team and are live-streamed on YouTube for public engagement. The process involves reviewing previous improvement tasks, celebrating successes, identifying shortcomings, and engaging in blameless problem-solving to enhance future releases. The retrospectives are documented for easy reference, with an assigned owner to manage and triage improvement items. GitLab also encourages asynchronous retrospectives to accommodate its globally distributed, remote team, allowing for flexible participation and documentation of insights. This practice not only aids in refining development processes but also strengthens the cultural fabric of the team by promoting continuous learning and improvement, key attributes of a robust Agile team.

Refactoring code is akin to maintaining a tidy kitchen; it simplifies the design of software without changing its behavior, which enhances maintainability and readability, thereby accelerating Agile delivery. This process is essential for preventing code rot, such as bad dependencies and duplicate code, and supports a culture of shared responsibility, trust, and collaboration among developers. By adopting practices like incremental refactoring and test-driven development, teams can systematically improve their codebase, making future development easier and more efficient. While refactoring requires a time commitment, the long-term benefits to developer productivity and system longevity are significant, as it allows for seamless integration of new features and simplifies QA and debugging processes. Embracing these techniques as part of an Agile mindset can transform team culture, promoting innovation and adaptability to change.

A new partnership between GitLab and Dotscience is announced, aiming to integrate DevOps tools and processes into the AI/ML ecosystem to achieve MLOps. This collaboration offers a comprehensive platform by merging source repository, issue tracking, and continuous integration, allowing users to build, train, deploy, monitor, and reproduce data, code, and models while collaborating on notebooks. The integration enables applying GitLab's familiar merge request workflow to AI and ML projects with Dotscience and customizing Docker images for deploying ML models to production. Users can start experiencing this integration by signing up for a free trial at Dotscience, which is available on various platforms, including on-premises and multi-cloud configurations.

Athlinks, known for its comprehensive suite of time management solutions for racing events, transitioned from using Jenkins to GitLab Ultimate to enhance transparency and communication across its teams. Previously relying on various Agile planning tools that did not meet their needs, Athlinks sought a system that could provide visibility and input from other business departments. After initially migrating to GitLab CE from GitHub in 2015 and later adopting GitLab EE in 2018, they conducted a proof of concept with GitLab CI, which led them to move to GitLab Ultimate, significantly reducing the number of jobs from 300 in Jenkins to less than 40. This shift facilitated a 50% reduction in time and code efficiency, improved collaboration among marketing, DevOps, and engineering teams, and ensured all issues, code, and CI pipelines were centralized, providing a single, comprehensive view of projects.

A Frontend Engineer at GitLab's Monitor::Health Group shadowed an on-call Site Reliability Engineer (SRE) to gain insights into incident response and improve empathy for users. During this experience, two incidents were observed: a crypto miner issue and a performance degradation on GitLab.com's applications. Key learnings included the importance of alert systems that detect deviations from the norm, the frequent non-critical nature of alerts, and the tools used by SREs, such as PagerDuty, Slack, Grafana, and Kibana. The engineer also recognized the potential risks of GitLab.com as a single point of failure and proposed enhancements like multi-user editing and additional webhooks in GitLab. The experience highlighted the complex nature of SRE roles and the potential for GitLab products to address operational challenges, with plans for future team expansion and product development to support these workflows.

GitLab's remote work culture, involving over 1,000 team members across 65 countries, emphasizes flexibility and personal comfort by having no physical offices, which allows employees to work from any location they prefer. This environment fosters unique and often humorous experiences during video calls, such as unexpected appearances by pets, children, and even wildlife, as recounted by team members like Tim Zallmann and others. The company embraces these occurrences, encouraging employees to bring their whole selves to work, which includes sharing their personal surroundings and interactions during meetings. This approach is highlighted by initiatives like virtual happy hours and serendipitous meetups, reinforcing a sense of community and camaraderie among remote workers. GitLab's culture promotes the integration of personal life with professional duties, leading to a vibrant and inclusive atmosphere where employees can express their individuality.

GitLab announced its intention to migrate key observability features such as custom metrics, logging, tracing, and alerting from a proprietary codebase to its open-source Core product, aiming for completion by early next year. This decision aligns with GitLab's open-core product philosophy, which emphasizes providing developer-focused tools in the Core to encourage wider usage and feedback from the developer community. By moving these features to the Core, GitLab seeks to enhance its end-to-end DevOps solution by addressing previous limitations in monitoring and observability within its open-source version, especially for users deploying applications on Kubernetes clusters. The company encourages community contributions to improve the observability tools, highlighting integration with Prometheus for custom metrics, live log streaming for logging, and the potential for deeper integration of tracing capabilities with Jaeger. GitLab is particularly interested in contributions that enhance the user experience in observing and managing the health of deployed applications, as well as in refining alerting capabilities. The initiative underscores GitLab's commitment to fostering collaboration and innovation within its developer community to further develop its observability suite.

Building enduring software products and companies requires a shift from output-focused metrics to outcome-driven strategies, emphasizing the importance of understanding true value through a value exchange system. The text uses the fictitious company Acme Co. to illustrate the challenges faced by tech firms, highlighting that despite high productivity metrics, the company's revenue growth is stagnating due to delayed delivery of promised capabilities. This stems from a misalignment between sales expectations and R&D's value delivery pace, exacerbated by a focus on output metrics rather than customer-centric outcomes. Acme's working group employs the Improvement Kata from the Toyota Production System to diagnose issues, discovering that a reduction in lead time could significantly boost revenue growth. The narrative underscores the necessity of optimizing both value delivery and capture and the potential pitfalls of relying solely on productivity analytics without considering the broader value exchange dynamics.

A working group, inspired by the Toyota Production System, set out to improve their company's value delivery by reducing waste, particularly time spent waiting, and implementing practices like pair programming to enhance efficiency. Despite achieving a 57% reduction in lead time and significant improvements in flow, the expected revenue growth did not materialize, prompting further investigation. They discovered that the focus on feature delivery had led to mounting technical debt and increased defects, which hampered new value creation. After convincing leadership to address technical debt and defects, the team improved efficiency but still struggled with revenue growth. This led them to explore throughput accounting, a concept emphasizing value creation over cost-cutting, which helped them realize the need to align their value delivery and value capture systems better. By focusing on increasing throughput, minimizing inventory, and reducing operating expenses, the group aimed to enhance the company's financial health, learning from the Theory of Constraints introduced by Eliyahu Goldratt.

Acme, facing financial challenges and potential layoffs, discovered systemic flaws in its operations, primarily due to a misalignment between what was being built and actual customer needs, known as the "tree swing problem." The company realized that its departments were working in silos, focusing on local optimizations without clear connections to overall business outcomes. This led to the adoption of throughput accounting, replacing traditional cost-based models to better measure success by outcomes rather than outputs. Acme also embarked on implementing a decision support framework using machine learning and natural language processing, allowing for more informed decision-making by analyzing data from existing tools without disrupting business processes. Through these efforts, Acme shifted to a continuous innovation framework, emphasizing shorter, more effective feedback loops, and aligning teams around customer needs and business outcomes. This transition helped Acme move from a reactive to a proactive, throughput-driven organization, significantly improving its financial trajectory and decision-making processes, while highlighting the importance of maintaining discipline to avoid reverting to old habits.

Breaking down silos between development and operations teams often leads to cultural shifts, creating a need for effective group development to maintain Agile methodologies. Bruce Tuckman's stages of group development—forming, storming, norming, and performing—offer a framework for rebuilding teams, fostering collaboration, and enhancing productivity. In the forming stage, teams set goals and establish interpersonal norms; in the storming stage, conflicts arise as trust is built. The norming stage sees the resolution of differences and a stronger team identity, while the performing stage allows teams to function autonomously and focus on continuous improvement. Agile leaders, particularly Scrum Masters, play a crucial role in guiding teams through these stages by fostering communication, trust, and a shared vision, ensuring that the team remains cohesive and productive despite changes in leadership or structure.

AWS re:Invent, a major event for tech enthusiasts, was filled with exciting announcements, particularly regarding AWS's impact on companies like GitLab, Goldman Sachs, and Verizon, each sharing their transformational journeys with AWS technologies. Notable developments included the launch of Amazon Fargate for Amazon EKS, which promises to simplify container management, and the introduction of the next generation of AWS Graviton processors, supported by GitLab. The event highlighted how companies, including Wag!, are leveraging AWS and GitLab to automate processes and integrate security earlier in the development lifecycle, enhancing efficiency and reducing costs. Additionally, GitLab's participation in AWS GameDay and their creative booth at the event underscored their active role in the tech community, while discussions on transformation and reinvention were central themes throughout the conference.

The GitLab Q4 Hackathon, held on November 13-14, set a new record with 109 Merge Requests (MRs) submitted, of which 79 were merged by November 25th. The event saw active participation from new community members and featured tutorial sessions with recordings available online. A notable contribution was an Epic created by Senior Frontend Engineer Winnie Hellmann, which provided manageable issues for participants and resulted in nearly 30 MRs. Another highlight was Utkarsh Gupta's MR, which focused on using gender-neutral pronouns in the GitLab handbook. Gupta was also the grand prize winner, having merged 28 MRs, breaking the previous record of 13. The event awarded GitLab branded tech organizers to participants with merged MRs, with Lee Tickett securing second place with 15 MRs. The next Hackathon is scheduled for February 12-13, 2020, and those interested in contributing can start by visiting the Contributing to GitLab page.

GitLab's public bug bounty program, launched a year ago, has been instrumental in enhancing the security of its product by engaging a global community of security researchers. Over the past year, the program received 1,378 reports from 513 researchers, awarding a total of $565,650 in bounties for valid vulnerabilities, which has kept GitLab's engineers and security team vigilant and innovative. In a recent bug bounty contest held from October to November, GitLab celebrated the achievements of contributors in various categories, including most reputation points, best written report, most innovative report, and most impactful finding, with winners being awarded custom mechanical keyboards. GitLab remains committed to continuously improving its bug bounty program and values the community's input to further advance its security measures.

GitLab is extensively used by marketing teams for project management, exemplified by the Just Commit campaign, which demonstrated the platform's capabilities in organizing and unifying complex projects with numerous stakeholders through features like issues, epics, and labels. The campaign utilized ancestor and child epics to manage goals, stakeholders, and timelines, ensuring collaboration and accountability. GitLab's issue boards allow teams to visualize and organize tasks effectively, with strategic placement of boards at different levels depending on the team's needs. Labels on issue boards help track the status of tasks, and the platform supports Agile methodologies through tools like milestones and burndown charts. While GitLab is primarily used by engineers for Agile code development, its functionality as a project management tool is adaptable across various departments, fostering collaboration and efficiency in diverse workflows.

GitLab has been recognized as the world's most productive remote team by focusing on efficiency, transparency, and results-oriented practices. The company empowers its remote employees to manage their own time and work autonomously, eliminating the need for a traditional office environment and fostering a culture of self-service through extensive documentation and asynchronous communication. Meetings are approached with flexibility, allowing team members to prioritize relevant content and encouraging documentation over real-time discussions, which promotes considered responses and minimizes unnecessary interruptions. GitLab prioritizes results over vanity metrics, ensuring that Objectives and Key Results (OKRs) drive productivity, and encourages minimal iteration to facilitate quick adjustments and foster innovation. By leveraging a comprehensive handbook for information sharing, the company minimizes repetitive inquiries and enhances productivity while maintaining an inclusive and innovative work culture.

GitLab is utilized beyond technical roles for collaborative project management, effectively addressing the common issue of siloed communications. The first part of a two-part series demonstrates how GitLab's structure facilitates collaboration for developers and non-developers alike, using an analogy of a layer cake to describe how its integrated tools streamline project workflows. This approach enables various teams, such as marketing, to manage complex projects by creating frameworks, assigning directly responsible individuals, and utilizing templates for repeated tasks. The marketing team, in particular, has developed a project management structure within GitLab that allows multiple teams to collaborate seamlessly, using epics and issues to track progress efficiently. By employing labels to indicate the stages of tasks, team members can maintain a global overview of projects, ensuring that tasks move smoothly from conception to completion. The article highlights GitLab's core values of collaboration and efficiency, suggesting that its tools are designed to enhance productivity across different sectors.

GitLab's all-remote work model fosters inclusivity and diversity by enabling the hiring of talent from a global pool, unconstrained by geographic location, which naturally mitigates biases and enriches the company's cultural tapestry. This approach aligns with GitLab's values of Collaboration, Results, Efficiency, Diversity, Inclusion & Belonging, Iteration, and Transparency (CREDIT), emphasizing the importance of hiring based on passion and results rather than location. By allowing team members to work from anywhere, GitLab supports individuals with varying needs, such as caregivers and those with physical disabilities, while bolstering local communities and reducing strain on metropolitan infrastructure. The model also offers cost savings by paying local rates, allowing GitLab to reinvest in its workforce as the company scales. Overall, the all-remote strategy enhances empathy and collaboration across time zones, creating a workplace where career advancement is not limited by physical presence in an office.

GitLab's engineering internship program offers a fully remote experience that promises flexibility, global collaboration, and valuable learning opportunities, making it ideal for students and career changers. The remote setup allows interns to balance work and life effectively, with the ability to work from anywhere and adjust hours according to personal schedules. Interns at GitLab are integrated into the team as if they were full-time employees, participating in major projects, and benefiting from a diverse and supportive team environment. The company's values emphasize creativity, transparency, and efficient workflows, allowing interns to contribute ideas that can be implemented after collaborative discussions. Through this internship, individuals can develop essential skills and potentially secure full-time positions, as demonstrated by the author's transition from an intern to a full-time role within six months. Overall, GitLab's remote internships provide a comprehensive work experience that encourages growth and development in a supportive and inclusive setting.

The blog post discusses strategies for organizing multiple product teams within GitLab to contribute effectively to a single repository, emphasizing GitLab's structured workflow approach. GitLab organizes its product surface area into seven sections, each with stages and categories, managed by cross-functional teams called groups. These groups utilize standardized workflows through group-level issue boards and labels, aiding in issue management and project tracking. The post details the use of specific labels for ownership, workflow states, and issue types to ensure clarity and mutual exclusivity, facilitating seamless transitions between teams and enabling comprehensive project oversight. Additionally, the post explores alternative methods like the project approach and single issue tracker project, while highlighting the benefits of centralized label management and shared milestones for maintaining consistency across teams. The importance of aligning around a shared release cadence using milestones and the potential future enhancements for milestone flexibility are also discussed.

Cloud adoption is a strategic approach that organizations use to enhance scalability, reduce costs, and improve operational flexibility by utilizing online software and services rather than traditional on-premise setups. The process involves four key steps: assessment, planning, implementation, and optimization. During assessment, companies evaluate their cloud readiness, identify expertise gaps, and consider compliance regulations. Planning requires selecting suitable cloud models and potentially adopting a multicloud approach, which involves using multiple providers to meet diverse needs. Implementation focuses on collaboration, iteration, and adapting DevOps structures to support cloud initiatives. Optimization is an ongoing journey of continuous improvement, where organizations must keep pace with innovation by automating and integrating processes. Despite challenges such as internal barriers and the need for application modernization, cloud adoption is crucial for maintaining competitiveness and leveraging the benefits of the cloud economy.

GitLab emphasizes the importance of its community in fostering engagement and learning through meetups, with a global network of 36 groups and over 7,700 members. In 2019, these organizers hosted more than 100 events, and the company anticipates growth in these numbers for 2020. GitLab encourages participation by inviting community members to attend meetups or start their own, offering support for new organizers. The December meetups span various locations, including México City, Wellington, Abraka, Bauchi, São Paulo, Minneapolis, Chicago, Geneva, and Warri, covering topics like GitLab introductions, version control, project management, CI/CD, GitOps, and AutoDevOps. These events provide opportunities for hands-on learning, networking, and sharing experiences within the DevOps community.

GitLab's strategy for the next three years focuses on enhancing its Dev section to become the world's leading product creation platform by improving efficiency, collaboration, and user experience across the DevOps lifecycle. With a market estimated to grow from $3 billion in 2019 to $7.5 billion in 2023, GitLab aims to increase its market share by offering industry-leading solutions in source code and code review, which serve as the initial entry point for users into the platform. The strategy emphasizes automated code reviews, improved value stream management, and accommodating various personas, including product managers and designers, to foster better cross-functional collaboration. GitLab also plans to address enterprise needs with enhanced compliance, governance, and workflow management features. Over the next year, GitLab will focus on making significant advancements in real-time collaboration, code review, project management, and enterprise readiness, aiming to provide a comprehensive and efficient DevOps platform that supports modern digital transformation demands.

OpenCores, a prominent community for Gateware development with over 300,000 members and more than 1,200 projects, is transitioning to GitLab to enhance IP core development and integrate innovative uses in scientific and electronic design communities. Gateware, defined as the intermediary layer between hardware and firmware, is crucial for digital designs, often involving FPGA and ASIC devices. OpenCores, maintained by Oliscience, supports the electronics design community by hosting a vast repository of open-source IP cores, promoting open standards, and fostering global collaboration. The community's move to GitLab, driven by its comprehensive DevOps and CI/CD capabilities, aligns with a broader industry push to simplify programmable logic development and implement Continuous Verification (CV) processes to ensure high-quality gateware. This transition is supported by strong ties with organizations like CERN and the European Space Agency, which are interested in improving gateware quality for industrial and commercial applications, highlighting GitLab's role in reducing design cycle times and enhancing code collaboration.

GitLab’s approach to updating its iconography emphasizes incremental changes through a minimum viable change (MVC) strategy, reflecting their commitment to efficiency and continuous improvement. With over 277 custom SVG icons and more than 800 FontAwesome icons, GitLab aims to create a consistent and strong visual identity while addressing the need for clarity and functionality in their user interface. The process involves revising documentation, categorizing icons for efficient use, and conducting multiple rounds of usability testing to ensure meaningful and proprietary visual representations, particularly for abstract concepts like epics and issues. Despite the challenges of working within constraints such as pixel grids and stroke weights, the iterative effort has been positively received, with ongoing plans to reduce reliance on external icon libraries, enhance documentation, and improve naming conventions. User feedback and collaboration are encouraged through the GitLab First Look program, allowing the community to play an active role in the evolution of GitLab's iconography.