How did you migrate without a freeze?

We ran old and new pipelines in parallel. For each service, we cut a feature flag, shipped the new pipeline and GitOps manifests, and used Argo Rollouts to route 10% traffic, then 30%, then 100%. Rollback was one `kubectl argo rollouts rollback` away. No big-bang weekends.

Why GitHub Actions over Jenkins?

We needed consistency, marketplace actions for security scanners, and fewer pets to patch. Actions gave us reusable workflows, per-repo configs, and simple org-wide policies. We kept a minimal Jenkins for edge workloads and retired it within 6 months.

How did you handle database migrations?

We templated `Flyway`/`Liquibase` steps in the pipeline and gated canary rollout until migrations were applied. For risky migrations, we used LaunchDarkly to decouple schema rollout from feature exposure.

What about AI-generated code and supply chain risk?

We treated it like any other code risk. Semgrep rulesets flagged insecure patterns, Trivy scanned images, and Cosign signed artifacts. We also diffed SBOMs in PRs and blocked unknown sources. The goal is guardrails for speed, not locks that force shadow IT.

Case-studies · Dec 11, 2025 · 10 minute read

From Jenkins Snowflakes to GitOps: The Platform Migration That Cut Lead Time by 92%

How we took a mid-market fintech from brittle pipelines and ticket queues to a sane, automated platform that devs actually like using.

Back to all posts

From Jenkins Snowflakes to GitOps: The Platform Migration That Cut Lead Time by 92%

Key takeaways

Implementation checklist