Back Issues/Search Home → Calendar → Archive → RSS → Subscribe → Services → Vendors → Current Issue → Popular → Newsletters (LinkedIn, Substack)

All issues › Volume 344, Issue 1 › IT News › Platform Engineering

Provisioning Went From Weeks to Hours. Here Is What Actually Changed

Platform Engineering, Tuesday, November 3rd, 2026

Lucid Motors cut Kubernetes cluster provisioning from weeks to hours by replacing ticket queues with a GitOps-based paved path.

Lucid Motors previously provisioned clusters through a chain of tickets with separate owners, which caused delays and inconsistent snowflake clusters.

The team replaced this with merge requests built around one values file of about twenty keys per cluster, from which Argo CD applications, IAM roles, DNS delegation and cost tags are derived.

Cluster API handles lifecycle, Terraform with Atlantis provides reviewable account-level changes, and a single Argo CD hub delivers the platform baseline to every cluster.

A self-service portal turns common requests into the same commits, and end-to-end cluster creation dropped from weeks to hours.

more →  ·  More from Platform Engineering →