Scalability

Half a million requests under 10 minutes with 1K concurrent connections

To demonstrate the scalability of the platform, see below the output of a simple load test with wrk.

Load test run with 1000 connections for a duration of 10 minutes

wrk -t50 -c1000 -d600s -s wrk.lua "http://localhost:3090/v1/api?csrf=..."

Each request validates the Authorization header, verifies that the user have access to perform a query on the given datasource, fetches the datasource metadata from Redis Cache or the database, if the cache is invalid, constructs and executes the SQL query against the MySQL database and returns the results.

wrk.lua script file content

Docker Stats after before the run

wrk output

Docker Stats after the run

Docker Resources - 6 CPU - 4GB RAM