Concurrency Budget
One request path, three limits: a thousand requests in flight at the API, fifty database connections, a hundred concurrent calls to an external provider. Set each limit and watch where the queueing lands. The lesson is uncomfortable and it is the whole page: a concurrency limit protects the thing you are calling, not you.
SIMULATED
No number on this page was measured. Wait times come from a standard multi-server queueing approximation, given the service times stated on each layer. The model reproduces the behaviour of real systems — work does not overlap on one core, a critical section serialises whatever fraction of the work it covers, a queue whose arrival rate exceeds its service rate grows without bound — but the figures belong to no real machine, runtime or workload. Read the shape of the curve and the direction of the change; never quote the milliseconds.
Predict first
Defaults, 600 requests per second, 40% of them touching the provider. Commit before the model runs.
At 600 req/s with the shipped limits, which layer saturates first?
Capacity at a layer is its limit divided by its service time. Do that division for each of the three before you answer.
Nothing below is shown until you commit. Deciding wrong and being told why is the exercise.