Latency SLA Budget

Problem

Compute how many decode tokens fit in a latency SLA: floor((slo_ms - prefill_ms) / decode_ms_per_token).

The signature is latency_sla_budget(prefill_ms, decode_ms_per_token, slo_ms). Return an integer.

Hint

Prefill and decode budgets are usually tracked separately in product SLAs.

case 1

in: [50,20,1050]

expected: 50

case 2

in: [50,20,50]

expected: 0

case 3

in: [100,5,200]

expected: 20

case 4

in: [0,10,55]

expected: 5

Starter codePython

def latency_sla_budget(prefill_ms, decode_ms_per_token, slo_ms):
    # Your code here
    pass

4 test cases · run your code against them in the editor