M3.5 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
The first question when picking a showback dimension is:
A. Which dimension has the best tag coverage
B. Which is cheapest to compute
C. Which the CFO prefers
D. What question the audience is asking
Show answer
Correct: D. Organisation answers “what do we spend”, teams answers “who spends it”, tags answers “on what”. Picking by coverage produces a report nobody asked for.
Q2
A resource shared by three teams is attributed by default as:
A. An equal split across owning teams
B. Full cost to the first team listed
C. Left unattributed until set by hand
D. Assigned to the account owner only
Show answer
Correct: A. Equal split is wrong when consumption is genuinely uneven, and usage-based attribution via tags is the alternative for those cases.
Q3
dimension_source distinguishes:
A. Provider derived from the resource’s region
B. Team from tag attribution rules
C. A cloud-native tag from an accepted Smart Tag
D. Calculated from actual cost rows
Show answer
Correct: C. It is how you answer “where did this chargeback number come from” when a team disputes it, and it lets a report filter to cloud-only for stricter attribution.
Q4
A tagging policy deriving one value across 100% of the estate means:
A. Attribution is now effectively solved for the org
B. Coverage went green while attribution stayed at zero
C. The policy needs a second tag key
D. Coverage should simply be recomputed
Show answer
Correct: B. Judge a policy by distinct values against real owners, never by coverage percentage. The easiest policies to write score best on coverage and attribute the least.
Q5
Tag coverage targets by maturity typically run:
A. Rising from a Crawl-stage baseline toward under 5% untagged at Run
B. 100% at every maturity stage
C. Fixed at 80% across the board
D. Provider-dependent, so the target varies from cloud to cloud
Show answer
Correct: A. Coverage is a trailing indicator of tagging discipline, and setting a Run-stage target on a Crawl-stage org produces a number nobody believes.
Q6
Unit economics requires a denominator that:
A. Grows in step with the infrastructure footprint
B. Is easy to collect automatically
C. Is supplied by the cloud provider
D. Moves independently of the cost being measured
Show answer
Correct: D. An infrastructure-derived denominator such as instance count moves with the cost, so cost-per-unit stays flat no matter what you optimize and the ratio cannot detect improvement.
Q7
The three unit-metric ingest paths are:
A. Manual, automatic and hybrid
B. Push API, CSV upload, Pull API
C. Webhook, cron, stream
D. Agent, API, export
Show answer
Correct: B. All three write into the same fact table via idempotent upsert, so mixing them during a migration does not double-count.
Q8
The Pull API’s SSRF guard works by:
A. Requiring an allowlist of pre-approved destination domains per organisation
B. Permitting only HTTPS on the webhook target
C. Resolving DNS then validating the resolved address, with redirect re-validation
D. Requiring mTLS from the receiving endpoint
Show answer
Correct: C. Resolve-then-validate is the load-bearing detail: validating the hostname before resolution is bypassable by a DNS record pointing at an internal address.
Q9
The unattributed residual is added to Summary and Trends but not to Teams or Tags because:
A. It is a provider-level figure only, with no underlying resource behind it
B. Showback views are cached daily
C. It would double-count the spend
D. Unattributed spend has no resource, so attributing it would mean inventing an owner
Show answer
Correct: D. RBAC-scoped views skip it for the same reason. Org totals reconcile with the invoice; showback attributes what can be attributed.
Q10
Switching an AWS org from Cost Explorer to a CUR 2.0 export most significantly:
A. Makes the historical costs amortized across the full commitment term rather than unblended
B. Lifts the 14-day resource-level limit, so historical per-resource showback becomes reconstructable
C. Removes unmatched rows from the unattributed residual
D. Excludes credits, matching the pre-credit invoice line
Show answer
Correct: B. It also ingests every invoice line-item type. A broken export fails the sync explicitly rather than silently degrading to the 14-day API.
What’s next
Back to Showback design.