M2.1 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
Which list contains only category values that live rules emit from Category()?
A. idle, rightsizing, orphan, compliance, sustainability, waste
B. cost, security, compliance, governance, performance, availability
C. idle, rightsizing, orphan, compliance, reliability, advisory
D. Per-provider and variable across AWS, GCP and Azure catalogs
Show answer
Correct: C. The registry emits 11 values: idle, rightsizing, schedule, orphan, discount, performance, compliance, reliability, security, governance and advisory. reliability is the largest of them (132 rules).
Q2
A rule produces a genuine $0 posture finding. Filing it under idle would:
A. Work fine; the category does not affect visibility at all
B. Raise the finding’s severity automatically on emission
C. Route the finding straight to the auto-remediation path
D. Cause it to be dropped by the low-savings suppressor
Show answer
Correct: D. Cost-driven categories must carry a real dollar figure, and cost-driven findings under $5 are dropped. Non-cost-driven categories exist so a genuine $0 finding survives, and orphan is explicitly exempt from the drop.
Q3
RC-004 and RC-006 no longer appear in the catalog because:
A. They were security rules and were moved to compliance
B. They were merged into RC-003, EC2 Rightsizing
C. They applied only to GCP and were dropped on v2
D. They were renamed RC-1004 and RC-1006 in the K8s block
Show answer
Correct: B. RC-003 absorbs the one-tier and two-tier paths, and RC-004 was removed from the auto-remediation allowlist at the same time. Consolidation is routine: RC-013 similarly merged into RC-097.
Q4
TierRates and SKURates differ in that:
A. TierRates carries commitment math; SKURates carries catalog-diff math
B. One is for AWS and the other is for Azure catalogs
C. TierRates is cached from the last pricing sync; SKURates is live
D. They are aliases for the same underlying rate lookup
Show answer
Correct: A. TierRates handles on-demand-versus-committed rate ratios; SKURates handles the rate delta between a current SKU and a target SKU. A savings number that defies intuition is usually a rule reaching for the wrong one.
Q5
A rule implements MetricsAware when it:
A. Emits metrics about its own evaluation runs
B. Needs CloudWatch, Cloud Monitoring or Azure Monitor data to evaluate
C. Runs on a metrics-driven schedule rather than on a fixed interval
D. Reports its findings to the metrics dashboard
Show answer
Correct: B. PricingAware is the sibling interface for rules that need per-resource monthly cost. Both are legacy injection interfaces: migrated rules read metrics and cost from the evaluation context instead.
Q6
The severity ladder runs:
A. P1 through P4, by response time
B. red, amber and green, by urgency
C. blocker, major, minor, trivial
D. critical, high, medium, low, info
Show answer
Correct: D. The timelines attached to them are an operational contract rather than aspiration: critical is today, high is this week, medium is this sprint, low is backlog, info is no action.
Q7
For a cost finding, which input decides the severity band?
A. The saving relative to org spend
B. The number of tags it carries
C. The age of the resource
D. The cloud provider
Show answer
Correct: A. Cost findings are scored on savings against an org-relative reference. Confidence and waste_ratio are low-weight tie-breakers that cannot cross a band alone. Posture findings are not scored: they keep the severity their rule declares.
Q8
The schedule category was carved out in a 2026 migration so that:
A. Schedules could then be auto-applied to matching resources
B. It could then be excluded from the organisation’s headline savings totals whenever reporting upward to leadership
C. Schedule rules could then skip the $5 low-savings floor
D. Scheduling-driven savings can be filtered as a first-class category rather than hiding inside idle and rightsizing
Show answer
Correct: D. The lever changed, so the category followed: these findings call for a recurring off-hours window rather than a one-shot change, and that distinction matters for triage.
Q9
A recommendation card’s evidence panel contains:
A. The cloud monitoring values that caused the rule to fire
B. The rule’s own source code and thresholds
C. The customer’s own written notes attached to the finding
D. A cost forecast for the coming month
Show answer
Correct: A. Evidence is what makes a recommendation auditable rather than an assertion, and reading it is the difference between applying a finding and trusting one.
Q10
The right response to a recommendation you cannot act on this quarter is:
A. Apply it anyway
B. Dismiss it permanently (Forever)
C. Dismiss it until next quarter
D. Ignore it
Show answer
Correct: C. Snooze is a dismissal with an end date (1 week, 1 month by default, 3 or 6 months). When it ends, the finding reopens. Only Forever is permanent, so keep it for “not applicable to us”.
What’s next
Back to The 684-rule library, explained.