Case Study: PicsArt achieves lower latency and 28% lower compute costs with Granulate

Fast-Growing PicsArt Utilized Granulate To Substantially Improve QoS While Lowering Compute Costs

PicsArt, the mobile social image editing application with 600+ employees and over 1B downloads, needed to support fast growth while processing large images and videos at scale in near real time. Its Kubernetes-based, on-premise Node.js application was latency sensitive, handling around 6K requests per second per node under strict SLA requirements, and the team wanted to improve performance without letting compute costs spiral.

PicsArt implemented Granulate’s real-time continuous optimization solution, which required no code changes or R&D effort. With Granulate, PicsArt reduced average latency by 12%, cut CPU utilization by 10%, increased throughput by 10%, and lowered compute costs by up to 28%, helping the company improve QoS while maintaining the same service levels.

PicsArt

Gerasim Hovhannisyan

Director of IT Infrastructure Operations