fix(open-data,hubs): align commitment discount eligibility with FOCUS spend and usage categories

[Roland Krummenacher (RolandKrummenacher)]

🛠️ Description

Fixes #2279.

x_CommitmentDiscountSpendEligibility reported reservation pricing and x_CommitmentDiscountUsageEligibility reported savings plan pricing. That is the reverse of the FOCUS CommitmentDiscountCategory definition the toolkit uses everywhere else: a reservation is a Usage commitment (you commit to a quantity of usage) and a savings plan is a Spend commitment (you commit to an amount of money).

IngestionSetup_v1_2.kql contained both mappings ten lines apart — it assigned CommitmentDiscountCategory correctly, then assigned the eligibility columns the opposite way, under a // TODO: Check this to ensure it's correct comment. That TODO is now resolved.

What changed

Component	Change
src/scripts/Update-CommitmentDiscountEligibility.ps1	Mapping extracted into a new New-EligibilityRow helper and corrected. The change-summary key is now built from the row's own properties, so it can't drift from the CSV column order.
src/open-data/CommitmentDiscountEligibility.csv	The two columns are swapped for all 92,624 meters. 36,287 rows change value; the 56,337 meters eligible for both are unaffected.
IngestionSetup_v1_0.kql, IngestionSetup_v1_2.kql	The spend column now reads spMeters, the usage column reads riMeters.
docs-mslearn/toolkit/open-data.md	Corrects the two column descriptions and the sample rows, and states the FOCUS reasoning.
src/powershell/Tests/Unit/Update-CommitmentDiscountEligibility.Tests.ps1	Five new tests pin the direction and the published column order.

The exclusion of reservation price rows from their own eligibility (and x_SkuPriceType != 'ReservedInstance') travels with riMeters to the usage column: it describes what an on-demand charge could be covered by, not the commitment row itself.

The best practice query in docs-mslearn/best-practices/compute.md already used the correct mapping and is unchanged. data-dictionary.md and the Copilot Studio schema-reference.md also already described the columns per FOCUS, and now match the data.

🧪 Verification

Corrected CSV values checked against the live Azure Retail Prices API (api-version=2023-01-01-preview, meterRegion='primary') for a stratified sample of 12 meters covering all three value combinations — 12/12 match, including the 00020329-e657-5687-9d1a-0be9876e5116 meter from the issue and the two other sample rows in open-data.md:

MeterId                                API:RI    API:SP    | CSV Spend     CSV Usage     match
00020329-e657-5687-9d1a-0be9876e5116   True      False     | Not Eligible  Eligible      OK
00003b45-e996-5b04-b673-a2db710f9237   True      True      | Eligible      Eligible      OK
0003c425-b1df-551f-a855-9c95a08cb4ae   False     True      | Eligible      Not Eligible  OK

Test runs:

• Invoke-Pester ./src/powershell/Tests/Unit/* — 2397 passed, 0 failed (4 pre-existing skips)
• ./src/scripts/Test-PowerShell -Lint — 3587 passed, 0 failed
• HubsKqlOperators.Tests.ps1 + KqlJoinKinds.Tests.ps1 — 587 passed, 0 failed

📝 Note for #2246

Draft PR #2246 replaces the derived eligibility computation in both ingestion transforms with a lookup against the open data table. It touches the same lines, so it will conflict here. Once the CSV is corrected, that lookup inherits the corrected values — but its x_SkuPriceType != 'ReservedInstance' guard should move from the spend column to the usage column, as it does in this PR.

✅ Checklist

• Changelog entries added under both FinOps hubs and Open data
• Unit tests added for the corrected mapping
• ms.date updated on both modified docs-mslearn files

🤖 Generated with Claude Code

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.