Skip to content
Zenta
Menu

Success story.

Parallel Transaction Scaling

The company was experiencing peaks of up to 100,000 simultaneous transactions during peak season, which caused latency and occasional outages on its e-commerce platform.

Image for the Parallel Transaction Scaling case

We take on any challenge

At Zenta there are no obstacles we cannot overcome.

  • Challenge

    Ensure availability and low response time during traffic peaks, without increasing fixed infrastructure costs.

  • Solution

    Architecture based on horizontal Cloud Run and Pub/Sub to decouple critical components. We configured automatic scaling based on CPU and latency, and implemented caching in Memorystore. All managed with Terraform.

Results from our management

  • 99.9% uptime during the last Hot Sale.

  • Average latency <200 ms at peaks of 150,000 TPS.

  • 30% savings in infrastructure costs compared to previous design.