DevOps Transformation: Stories of Role Integration and Work Consolidation

In the context of DevOps, the principle "who develops, who operates" is often misunderstood as merely colocating developers and operations staff, which this article argues is insufficient.

1. DevOps transformation "queue‑jumping" stories

It recounts three narratives: (a) Xiao Ming, an operations specialist, was moved into the development team after the operations department was dissolved, yet his workload remained dominated by production support, leaving little time for development tasks; (b) DBA Xiao Cui was reassigned to delivery teams, becoming over‑burdened by multiple team requests and losing opportunities for professional growth; (c) DBA Da Peng, newly hired, was placed in a delivery team where his DBA duties were underutilized, leading him to perform unrelated project work and eventually resign.

2. Rethinking work integration in the DevOps era

The article questions which tasks should be integrated and which should remain separate, noting that traditional fine‑grained division of labor works well for repetitive manufacturing processes but not for software delivery, where each requirement is unique and knowledge transfer across roles introduces friction and errors.

It advocates for small, cross‑functional teams (ideally no more than seven members) that possess end‑to‑end delivery capability, reducing handoffs and external dependencies while acknowledging that certain specialized skills (e.g., DBA, infrastructure) cannot be fully absorbed by every team.

3. Key to solving work integration problems

The core is to identify repeatable versus non‑repeatable work. Repeatable tasks (infrastructure, deployment, DBA operations) should be standardized, automated, and possibly centralized; non‑repeatable tasks (requirements analysis, development, testing) should stay within autonomous delivery teams.

Practical suggestions include documenting deployment and incident‑handling procedures, automating them, empowering delivery teams with self‑service scripts, and consolidating expert resources to serve multiple teams without creating bottlenecks.

4. Conclusion

Organizational reshuffling is inevitable, but without careful analysis of which work can be standardized and which requires dedicated expertise, the intended efficiency gains of DevOps may be lost. The ultimate goal remains continuous delivery and improved overall delivery efficiency through thoughtful integration and separation of work.