Scaling GitOps With Release Promotion
Cloud Native Now, Tuesday, October 20th, 2026
GitOps that works at five clusters breaks down at hundreds; staged promotion with tracking and verification restores confidence.
Jesse Suen explains that a merge-syncs-everywhere GitOps design becomes a fleet-wide event with no gate once teams run around fifty or more clusters.
He cites a platform team with over 200 single-tenant production clusters that relied on one replica environment, where validation took weeks.
Synchronization tools distribute config but do not track which version runs where, what checks it passed, or where it can go next.
He advocates staged promotion through smaller cluster groups, with verification history and the ability to halt bad releases.