Incident response and capstones · Lesson 2 of 12 · 9 min

Investigate inference errors

Inference error triage distinguishes rejected inputs from runtime failures. Group by status and request ID before choosing a retry or configuration change.

Open the lab Course map

Mastery 0/3Run itExplain itApply it

learner@clairos:~ $

↑ history · Tab complete · Ctrl+C cancel · Esc, Tab to leaveExit 0

Your objectives

Hints are always available. Wrong turns never cost XP.

1
Read the case evidence
Read /opt/ai-lab/cases/aiops-incidents-capstones-002.json. Separate the recorded facts from your assumptions.
Hint & worked example
Inference error triage distinguishes rejected inputs from runtime failures. Group by status and request ID before choosing a retry or configuration change.
The complete case JSON is read successfully.
```
cat /opt/ai-lab/cases/aiops-incidents-capstones-002.json
```
Common mistake to avoid: Treat 98 validation errors as transient capacity failures.
2
Isolate the deciding signal
Select the most_common_error field from the case JSON. What does it establish—and what does it not establish?
Hint & worked example
Fix the missing model field driving most errors before retrying unchanged requests.
A successful JSON query isolates most_common_error.
```
jq .most_common_error /opt/ai-lab/cases/aiops-incidents-capstones-002.json
```
Common mistake to avoid: Using a different case or an unrelated signal as the justification.
3
Record your operational decision
Choose the justified code: fix-client-model-field, change-without-evidence, or ignore-the-signal. Save that one code followed by a newline in /home/learner/lab/aiops-incidents-capstones-002-decision.txt.
Hint & worked example
Fix the missing model field driving most errors before retrying unchanged requests.
The decision artifact contains the correct code, not just a matching command in history.
```
printf '%s\n' 'fix-client-model-field' > /home/learner/lab/aiops-incidents-capstones-002-decision.txt
```
Common mistake to avoid: Treat 98 validation errors as transient capacity failures.

Understand the pattern

Inference error triage distinguishes rejected inputs from runtime failures. Group by status and request ID before choosing a retry or configuration change.

Interpret the evidence

Fix the missing model field driving most errors before retrying unchanged requests.

A tempting mistake

Treat 98 validation errors as transient capacity failures.

All names, measurements, prices, versions, and thresholds in this exercise are fictional training data. The evidence supports this bounded scenario, not a claim about a live provider.

Worked example and expected result

$ cat /opt/ai-lab/cases/aiops-incidents-capstones-002.json

{
  "fixture": "Fictional training data; not a live measurement.",
  "errors_400": 98,
  "errors_503": 2,
  "most_common_error": "missing model"
}

Exit status: 0. The command completed successfully in the training fixture.

$ jq .most_common_error /opt/ai-lab/cases/aiops-incidents-capstones-002.json

"missing model"

Exit status: 0. The command completed successfully in the training fixture.

$ printf '%s\n' 'fix-client-model-field' > /home/learner/lab/aiops-incidents-capstones-002-decision.txt

(no output; exit 0)

Exit status: 0. The command completed successfully in the training fixture.

Try a variation

Name the assumption that would have to change for “fix-client-model-field” to stop being the justified action. Do not silently replace missing evidence with a guess.

Explain it

What does the evidence in “Investigate inference errors” justify?Treat 98 validation errors as transient capacity failures.Make the change now and gather the missing evidence afterward.Fix the missing model field driving most errors before retrying unchanged requests.

Apply it elsewhere

A colleague concludes: “Treat 98 validation errors as transient capacity failures.” What should you do?Treat this fixture as a live measurement.Reject that conclusion; return to the case evidence and its stated limits.Accept it because the training command completed.

Keep the momentum.

Complete the lab, then explain and apply the pattern.

Next: Diagnose retrieval regression Try an incident drill

Anonymous lesson and mastery progress lasts for this session. Sign in to keep synced lesson XP between sessions; mastery stays device-local. No speed bonus, lost lives, or hint penalties.

Keep learning

Case methodology: evidence, limits, decision