JYSK chose to implement Kubernetes as part of the Unified Commerce initiative across 3,400 edge locations in order to drive uplift across several metrics, including operational efficiency, stability, and customer experiences. Due to the size and scope of the project, the team needed to ensure cost-efficiency and long-term scalability. The team considered its computational resources such as processing power and memory as well as the widespread nature of the deployment, which left nodes vulnerable to local power outages and security breaches, posing a significant hurdle to cluster management.

JYSK’s initial strategy involved using a lightweight Kubernetes distribution (K3s), combined with GitOps practices and the Cilium CNI. While this setup worked during early testing, it quickly exposed critical scalability and operational issues and they faced significant challenges.

The sheer volume of updates and patches required for in-store clusters became unmanageable. Individual updates were not overly complex, but the frequency and scale of patches threatened to overwhelm the infrastructure. Centralized traffic routing introduced performance bottlenecks and pulling container images through a shared cache further complicated the architecture. JYSK was also constrained by the diversity of its store network environments which increased the difficulty of centralized management.

The POC was paused before fully reaching Day 2 operations as the infrastructure would cause long-term complexity and operational risk. The team needed a solution that was scalable, secure, and easy to automate while still fitting the realities of edge deployments, such as limited bandwidth, local power outages, and high hardware variability.