Retrospective: We Ditched Jenkins and Saved 30% on CI Costs Using GitLab 17 and Pulumi 3.150 for 2026 Teams

Our Jenkins Pain Points

By late 2025, our engineering organization supporting 12 distributed product teams (the core of our 2026 planning cohort) was drowning in Jenkins maintenance. We ran 45 self-managed Jenkins controllers across 3 regions, with over 200 plugins to support legacy pipelines, static agent pools, and custom Groovy scripts. Key issues included:

Idle agent waste: Static agent pools sat unused 40% of the time, driving up cloud compute costs

Plugin hell: Conflicting plugin versions caused weekly outages, requiring 2 full-time engineers to maintain

Pipeline portability: Jenkins pipelines were tied to our legacy on-prem controller, with no native support for hybrid cloud workloads we planned for 2026

Cost opacity: We couldn't attribute CI costs to individual teams, making it impossible to optimize spend

Why GitLab 17 and Pulumi 3.150?

We evaluated 6 CI platforms in Q4 2025,…