Flood is a load testing platform that lets you run globally-distributed performance tests with your favorite open source tools, including JMeter, Gatling and Selenium. Scale out your flood load tests for maximum concurrency and throughput at any given time. We’ll take care of the infrastructure and provide aggregated, real-time reporting. Their distributed grid infrastructure was built with a “shared nothing” architecture that lets us scale horizontally beyond the capabilities of any other load testing service on the market today. Whether you need to load test a single URL, simulate realistic browser behavior with Selenium, or execute large concurrency and volume with JMeter and Gatling, Flood provides a simple and affordable platform for scaling load tests on demand.

Customer Success Story: Flood.io

Flood.io is a distributed load testing platform that supports testing with open source tools like JMeter, Gatling, and Selenium. Their platform collects and aggregates real-time data from the tests conducted by their customers in InfluxDB Cloud.

Read Case Study   Watch Webinar

The Challenge

Flood IO’s first-generation design used Elasticsearch. Each grid (infrastructure that the tests run on) would operate in a cluster of 30 nodes. Each node ran Elasticsearch and they would pull that information from those grids. This early design presented several challenges:

  1. It was difficult to scale horizontally since cluster dependencies would impact grid node startup time.
  2. Although Flood IO was pulling data from the Elasticsearch clusters in real time, the data wasn’t visible to the user as such.
  3. They were using the full-text search database Elasticsearch to store time series data, which is not that database’s intended use.

The Result

The new design using InfluxDB Cloud helps Flood IO’s solution achieve its design goals of being a distributed, loosely coupled, shared nothing, load test platform. The system processes their cue asynchronously without any delays and runs all the points up into InfluxDB Cloud, which processes around 60 write points/sec.

Scroll to Top