Kubernetes platform for a logistics operator

The client is a Dutch logistics operator whose track-and-trace systems ran across 14 hand-built Kubernetes clusters. These systems are business-critical: customers and operations depend on live tracking.

The clusters had diverged over time into snowflakes, and the operational load of keeping them all running and upgraded fell on a stretched platform team.

Track-and-trace ran on a sprawl of 14 hand-built clusters that had diverged over time, each with its own quirks. No two were quite the same, so nothing could be done once and applied everywhere.

Every upgrade was risky because no two clusters were quite the same, and the operational load fell on a stretched team. Routine maintenance had become a source of both toil and anxiety.

The operator needed one consistent, governed platform without a disruptive migration of critical tracking workloads. Consolidation could not come at the cost of live tracking.

We standardised on a single governed platform with GitOps delivery, so cluster state is declared in version control and drift disappears. Declaring desired state, rather than hand-tuning clusters, is what ends the snowflake problem.

Automated, tested upgrades replaced the risky manual ones, making patching routine rather than an event. The fear went out of maintenance because the process became repeatable and reversible.

Workloads migrated cluster by cluster, so tracking stayed live throughout the consolidation, and the shared platform cut the operational surface the team had to carry. Fewer, consistent clusters meant far less toil.