Electricity bills: the due-date rush
The last days before the bill due date send payments and balance lookups through one Postgres. Stay fast and keep paying through a slow billing service and a lost instance, inside the budget.
ClosedWeek 2026-W43
The brief
The story
The last days before the due date are the busiest of the month for a state electricity utility's app. Customers check a balance, then pay it, and an acknowledged payment must never be lost. Every request ends in one Postgres, the source of truth for accounts.
The target
- Load: 2,800 requests a second for 8 seconds, nine in ten of them balance lookups and one in ten payments.
- Latency: p99 within 100 ms.
- Budget: $1,650 a month at this load. Half the budget or less earns full cost marks; one and a half times it earns none.
- Faults: two 6-second drills, each with its fault landing 2 seconds in. In one the billing service runs twice as slow for 3 seconds; in the other one service instance crashes. A drill passes when under 1% of requests fail and p99 stays within 200 ms.
What is scored
Throughput 30%, latency 20%, cost 15% and resilience 35%, out of 1,000,000. Resilience weighs most this week: a fast, cheap design that falls over on one lost instance cannot win. Every entrant runs the same arrival stream; its seed is published when the week closes so anyone can replay a score.
How to enter
Choose Open in Sandbox. The canvas starts from the palette and starter design this challenge is built on. Build, test it here, then enter. You get five scored entries.
The scenario
- Target load
- 2,800 requests per second, steady
- p99 latency target
- 100 ms
- Monthly budget
- $1,650 a month
- Fault script
- slow on the first service, from 2 s for 3 s
- crash instance on the first service, from 2 s
- How it is scored
- Throughput30%
- Latency20%
- Cost15%
- Resilience35%
Your entry
Leaderboard
0 entrants · updated 28 October 2026 at 1:26 pm IST
Nobody entered this week
The next challenge opens on Monday at 00:00 IST.
Seed
Every entry ran with seed 4310. Replay a design against it to reproduce its score.