Case Study: PicnicAI automates Kubernetes rightsizing and improves reliability with PerfectScale by DoiT

PicnicAI automates 85–90% of Kubernetes workloads with PerfectScale

PicnicAI, a health-tech company that processes medical data for clinical trials, faced significant challenges managing its complex Kubernetes environment. With a lean engineering team, they struggled to manually right-size resource allocations for their highly variable data-processing workloads, leading to frequent out-of-memory errors, job failures, and operational disruption. They partnered with vendor PerfectScale to address these reliability issues and automate optimization.

PerfectScale provided automated, workload-aware rightsizing and granular visibility into PicnicAI's Kubernetes clusters. With support from DoiT's technical teams, PicnicAI implemented a phased approach, using PerfectScale's policies to safely automate resizing. The results were substantial: automated resizing now covers 85-90% of workloads, reliability greatly improved, and the infrastructure team reduced time spent on management by approximately 50%. This eliminated over-provisioning for cost savings and freed engineers to focus on higher-value projects.

PicnicAI