Traceability in DevOps : Expectations, Challenges and Opportunities

Sammanfattning

Background: DevOps practices have become central to modern software development, enabling rapid delivery through continuous integration and deployment. At the same time, software traceability remains important for accountability, coordination, compliance, and quality assurance. However, maintaining traceability in fast-paced DevOps environments is challenging, and existing research has largely focused on plan-driven development or technical solutions without fully capturing current industry practice. Objectives: The objective of this thesis is to understand how software traceability is perceived, implemented, and sustained in contemporary DevOps environments. The study investigates current practices, challenges, and perceived impacts of trace ability on software quality, release cycles, and collaboration, as well as the role of tooling and automation in supporting traceability. Five research questions address: (1) current perception and implementation, (2) challenges in integration, (3) impact on delivery outcomes, (4) critical success factors for tool integration, and (5) the role of automation in feedback loops. Methods: A sequential mixed-methods approach was adopted. First, an online survey was conducted with 90 practitioners working in DevOps-related roles to capture broad patterns and perceptions. This was followed by 20 semi-structured interviews to explore experiences, workflows, and contextual factors in greater depth. Quantitative data were analysed using descriptive statistics, while qualitative data were analysed through thematic analysis. Results: The results show that traceability is widely regarded as important in DevOps, but is only partially and unevenly implemented in practice. Key challenges include high maintenance effort, fragmented toolchains, and inconsistent adoption across teams. Traceability is perceived to have a positive impact on software quality, collaboration, and compliance, while its effect on release speed is more mixed. Automation is seen as an important enabler, but not a complete solution, as effective traceability also depends on clear ownership and alignment with everyday work practices. Conclusions: This thesis concludes that traceability in DevOps should be understood as a socio-technical practice rather than a purely technical feature. Sustainable traceability requires a balance between speed and discipline, supported by integrated tools, lightweight automation, and organisational commitment. The findings provide empirical insights that can inform both future research and practical improvements in DevOps environments.