Skip to main content
Every tier at once: edge cache, load balancer, services, cache, shards, and a queue of async work behind it all.

What to watch

Two thirds of the traffic never gets past the edge cache. Only the misses reach anything below, which is why everything behind it looks so lightly loaded at rest. Take the load to 4x and watch both halves. The sharded store saturates and the client sees errors straight away. The workers saturate too, but the queue swallows the excess, so the client is still told everything is fine while a backlog builds that takes hours to drain.

What it is made of

Under load

Measured by running this design through the simulator at multiples of its own offered rate of 600 requests a second. Twenty seconds of simulated time, one fixed seed, so the same numbers come out every time.
This design holds 1500 requests a second with 0% errors. At 1800 it is losing 19% of them.

Open this example

Load Full stack from the Examples menu and drag the traffic slider yourself.