Module 10 Knowledge Check
5 questions. Pass mark 4 out of 5. Answer every question before checking the answer key below, then retry after reading the feedback.
1. The first thing a trace must show when a run goes wrong is…
- A. Total cost
- B. The sequence of nodes, the state at each step, and the actual inputs and outputs
- C. The model version
- D. The user id
2. A workflow produced a wrong final answer. The efficient first check is…
- A. Rewrite the final prompt
- B. Walk the trace to find the first step where state became wrong
- C. Change the model
- D. Add retries
3. Traces of production runs commonly contain…
- A. Only metrics
- B. User content and tool inputs, so they need access control and retention limits
- C. Nothing sensitive
- D. Only errors
4. Recovery after a failed consequential action must first establish…
- A. The cost
- B. Whether the action actually completed before the failure
- C. The model used
- D. The node name
5. Idempotency keys on action tools help because…
- A. They speed execution
- B. A repeated call with the same key does not repeat the effect
- C. They reduce cost
- D. They improve routing
---
Answer Key and Explanations
Check these only after attempting every question.
1. B - The sequence of nodes, the state at each step, and the actual inputs and outputs
Without per-step state you are guessing at which node introduced the problem.
2. B - Walk the trace to find the first step where state became wrong
Fixing the last node when the error entered three nodes earlier just moves the symptom.
3. B - User content and tool inputs, so they need access control and retention limits
Observability data is a second copy of your users' data; govern it accordingly.
4. B - Whether the action actually completed before the failure
Blind retry can duplicate a sent message, a payment, or a record change.
5. B - A repeated call with the same key does not repeat the effect
They make retries safe, which is what allows automated recovery at all.
Practical Check
Apply this module to your own work: complete the module activity for *Debugging, Tracing, and Failure Recovery*, then write one sentence naming what your result shows and one naming what it does not.
Strong Answer Pattern
A strong answer names the task, the evidence used, the check performed, and the remaining limitation. It avoids "proved", "guaranteed", or "always" unless the evidence genuinely supports it.