Logistics · OKE Platform

A logistics operator rebuilt on OKE and shipped releases in hours instead of weeks.

A national logistics operator was held back by a monolith that took weeks to release. We built a container platform on OCI Kubernetes Engine, and the release cycle dropped from weeks to hours.

Logistics warehouse with shelving and forklifts representing a logistics operator
hours
release cycle, down from weeks
60%
fewer failed deployments
Peak
capacity that scales with demand
24/7/365
managed run after go live
The situation

Where they started.

The operator ran its routing and tracking platform as a single large application on virtual machines. Every release meant a coordinated outage, a manual deployment, and a long weekend of testing, so the business shipped changes only a few times a year. When demand spiked at peak season the whole application had to scale as one block, which wasted capacity and still left gaps under load.

They wanted to release small changes safely and often, scale the parts that needed it without scaling everything, and stop paying for headroom they only used a few weeks a year.

What we did

The work.

  • Mapped the monolith into services that could be deployed and scaled on their own.
  • Built a managed OKE cluster with separate node pools for steady and bursty work.
  • Set up a pipeline so a commit moves through build, test, and deploy without manual steps.
  • Added horizontal autoscaling tied to real queue depth and request load.
  • Wired observability into every service so a bad release is caught in minutes.
  • Trained the internal team to own the platform with our managed run behind them.
The architecture

The OCI architecture used.

The platform separated the things that change often from the things that stay stable, so releases got small and scaling got precise. The cost came down because capacity now follows demand instead of sitting idle.

LayerWhat we usedWhy
OrchestrationOCI Kubernetes Engine (OKE)A managed control plane so the team runs apps, not Kubernetes
ComputeMixed node pools with autoscalingSteady baseline plus burst capacity only when peak demand arrives
DeliveryAutomated build and deploy pipelineSmall frequent releases with rollback built in
ObservabilityMetrics, logs, and tracing per serviceA bad release is visible in minutes, not after a customer calls
The release calendar stopped being an event. Small changes now ship in hours, and peak season no longer means paying for idle capacity all year.
Related

Where this connects.

Solution

OKE

How we design and run container platforms on OCI Kubernetes Engine.

See OKE
Workload

App Modernization

Breaking a monolith into services that deploy and scale on their own.

See modernization
Service

OCI Managed Services

The run model that keeps the platform healthy after go live.

See managed services

Want an outcome like this?

We anonymise every client by sector, but the method behind these results is the same one we would bring to your estate. Book an assessment and get a written plan with options and a price.