M2.13 module quiz
Ten questions. 80% to pass (8 of 10). Open book, unlimited retakes.
Answers are collapsed under each question. Answer first, then check.
Q1
status and resolution_state are:
A. Two independent columns: what the customer claimed, and what was measured
B. The same underlying value stored in two entirely separate columns
C. Sequential stages of one process
D. UI and API views of one field
Show answer
Correct: A. They are allowed to disagree, and the disagreement is the point. A savings report built on status alone is wrong in both directions and confident about it.
Q2
status = applied with resolution_state = verified_unchanged means:
A. Verification has not run
B. The rule stopped firing
C. The customer acted differently than asked
D. The customer clicked and nothing changed
Show answer
Correct: D. This is the lie the two-column model exists to kill. In one enterprise audit it was the largest single line, being reported upward as captured savings.
Q3
A customer resized to t3.large when the recommendation asked for m5.xlarge. The verdict is:
A. verified_unchanged
B. verified_diverged
C. verified_adopted
D. abstained
Show answer
Correct: B. They solved the problem, differently. Calling it adopted overstates our contribution; calling it unchanged is false. Diverged findings still count toward the defensible savings figure.
Q4
target_spec is written once and never recomputed because:
A. It reduces the number of database writes
B. Instance types are immutable once chosen
C. It would break the recommendation’s audit trail
D. A moving target would make adoption unachievable
Show answer
Correct: D. Someone who resized exactly as asked would be re-evaluated at their new shape, given a new target, and measured against that. A reopen clears the target and baseline so the next window starts with a matching pair.
Q5
“The rule stopped firing” is not accepted as evidence of a fix because:
A. Rules fire intermittently by design
B. Rules can be disabled by the customer
C. A genuine fix, a narrowed IAM role, a renamed field, a discovery gap and a failed metrics fetch all produce identical silence
D. It would close recommendations far too quickly to be useful in practice, before the customer had any chance to act
Show answer
Correct: C. Verification re-reads the key the rule declared and requires it to have moved to the compliant value. Under a stopped-firing heuristic, a permissions regression gets reported as a compliance improvement.
Q6
A resize targeting a key with no read-side mapping produces:
A. verified_unchanged
B. verified_diverged
C. pending
D. abstained
Show answer
Correct: D. Reporting “we cannot see this” as “unchanged” would silently accuse the customer of not acting. Only three of the nineteen target-spec keys are first-class columns; the rest live in metadata with partial coverage.
Q7
A schedule recommendation is verified by:
A. The resource’s current power state
B. The uptime fraction against the frozen baseline
C. Whether the rule has stopped firing
D. The presence of any schedule attachment at all
Show answer
Correct: B. An adopted schedule leaves the resource running by day and stopped by night, so a 6-hourly scan samples status at four arbitrary points and it becomes a coin flip.
Q8
Per-resource commitment claims:
A. Verify on the month boundary
B. Verify once coverage flips
C. Always abstain
D. Verify from purchase_type
Show answer
Correct: C. A commitment is an account-level financial instrument, so attributing its adoption to a single resource is not something the data can honestly support. It abstains by rule, not by circumstance.
Q9
Verification covers roughly what share of the rule catalogue today?
A. A third
B. All of it
C. Two thirds
D. A tenth
Show answer
Correct: A. About 129 of ~387 rule ids have an authoritative lever. Closing the gap is a per-rule declaration, and a registry-walking test prints the coverage number so it stays visible and monotone.
Q10
A rule version is cut by:
A. Renaming the rule in the catalog
B. Changing the rule’s severity level
C. Changing a threshold, lookback window or condition
D. Recategorising the rule under a different category
Show answer
Correct: C. Declaration changes reach the customer through the registry at render time. Only a behaviour change increments, and it is manual: nothing verifies the author bumped it.