e-commerce / Retail
5 clusters, 525 services
Nemlig.com is Denmark’s largest and most successful online supermarket with a wide array of discount and premium convenience products. Nemlig has over 1,400 employees, serving thousands of households, 365 days a year.
Nemlig.com has a very large and robust engineering organization that is tasked with keeping the various components of a busy eCommerce website running smoothly; from secure payments to real-time product data, everything must be in check in order to withstand the massive load of thousands of daily transactions.
The engineers at Nemlig were previously using a Docker host for container orchestration. When the decision was made to implement Kubernetes for that purpose, there was an urgent need to reduce the time to value significantly and to unlock the many benefits of Kubernetes as fast as possible.
Nemlig has a multi-cluster setup with a lot of workloads being distributed throughout those clusters. Keeping track of those workloads can be tricky for a K8s novice, but even a seasoned SRE will suffer from context-switching and alert fatigue. The Nemlig team was simply unable to extract the value they needed out of Kubernetes fast enough, given their sprawling architecture and a huge pool of developers.
Komodor’s platform provided a comprehensive solution to Nemlig’s challenges, enabling them to manage their workloads effectively and efficiently, across different clusters and cloud providers.
The service overview feature allowed them to see all the services and their health status without having to go into each one individually, saving time and effort. Additionally, the platform’s real-time and historical data allowed the DevOps team to investigate events and drill down to understand the overall performance and reliability. The instant issue identification and correlation of events enabled them to pinpoint issues immediately and resolve them with a click.
The platform’s suggestions for problem-solving further enhanced their ability to manage their workflows efficiently. Moreover, the platform’s overview of nodes and suggestions as to what might be wrong helped Nemlig identify issues that were not previously apparent, resulting in significant time savings in figuring out the problem.