What to watch
The database handles 100 a second. Once waiting passes the 250ms timeout on the API, every request starts retrying three times, so the load triples onto a database that was already full. It is steady at 80 a second. Push to 85 and it starts losing most of its traffic; by 90 almost none gets through.What it is made of
Under load
Measured by running this design through the simulator at multiples of its own offered rate of 45 requests a second. Twenty seconds of simulated time, one fixed seed, so the same numbers come out every time.This design holds 90 requests a second with 0% errors. At 113 it is losing 100% of them.
Open this example
Load Retry storm from the Examples menu and drag the traffic slider yourself.