Challenge · Serving and cost · Premium
Route the cheap traffic
At Monday peak Northwind's primary model sits at its rate limit while the fallback idles. Find the bottleneck, route the traffic that does not need the large model, and keep the policy reasoning where it belongs.
- intermediate
- about 25 min
- 5 objectives
- par 9 steps
- Northwind Insurance
Mission workspace
This mission is part of premium, with every challenge and incident in the lab. The brief below is the whole problem; the box, the hints and the debrief are what you unlock.
Get full Premium access · ₹2,000 / $25
Every answer, concept and course, every paper breakdown, all published video lessons, hands-on FDE Lab missions, Premium PDF guides and companion files, the full practice-test bank and work-sample downloads. Referral Premium excludes guide PDFs and their companion files.
6 months · One payment · No auto-renewal
Study alongside free video lessons.
Brief
- Customer
- Northwind Insurance (synthetic customer; every metric and trace on this box is made up for the exercise)
- Situation
- Monday 09:10. The claims assistant's primary provider is at 100% of its rate limit and answering 429s with no fault anywhere; the fallback provider is at 2%. Northwind's platform lead wants to know why they pay for a fallback that does nothing, and whether a bigger plan is the only answer. Every case runs the same model for every turn, from classifying the message to reasoning about a replacement.
- Your role
- You are the FDE who owns the router. You have the console and the router's config file, which the console binds to the live router. You cannot raise the provider's limit today.
- Goal
- Name the bottleneck from the metrics, route a share of calls to the fallback, prove the primary is off its limit on the live system, and say which class of call must stay on the large model.
Objectives
- 1.Name the bottleneck (you commit to an answer)
- 2.Route the cheap call classes to the fallback
- 3.Show the primary off its limit: saturation under 80% and 429s under 2% on the live system
- 4.Name the call classes that must stay on the large model (you commit to an answer)
- 5.Check whether the fallback can carry the routed load (you commit to an answer)