fix(providers/aws): fail loud on unparseable RI cost fields

[cristim]

Problem

parseAWSCostDetails in providers/aws/recommendations/parser_ri.go silently swallowed strconv.ParseFloat errors for UpfrontCost, EstimatedMonthlyOnDemandCost, and RecurringStandardMonthlyCost, leaving the destination field at 0 with no log. An unparseable UpfrontCost would surface an all-upfront RI recommendation showing $0 upfront, a wrong money figure feeding effective-savings math and purchase decisions, with zero signal. This was inconsistent with parseCostInformation in the same file, which fails loud for the savings fields (COR-07 in docs/reviews/codebase-review-2026-06-10.md).

Fix

• parseAWSCostDetails now returns an error when a present cost string is unparseable, matching the parseCostInformation pattern; absent (nil) fields keep their existing behavior.
• parseRecommendationDetail propagates the error with context; the existing caller chain (parseRecommendations) logs a warning and skips the affected recommendation detail, so a malformed CE response drops that detail loudly instead of fabricating zeros into money math.

Test evidence

• New regression test TestParseRecommendationDetail_MalformedCostFields covers all three malformed fields plus a well-formed sanity case. Confirmed FAILING with the pre-fix code (stashed fix: 3 subtests fail with "An error is expected but got nil") and passing after.
• go test ./recommendations/ in the providers/aws module: 335 passed.
• go build ./... clean in both the root module and providers/aws.

Closes #1171

[cristim]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in: 58 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: aa3ae7cb-683d-44ba-9558-12e27e52c2b6

📥 Commits

Reviewing files that changed from the base of the PR and between f692085 and 23d6415.

📒 Files selected for processing (2)

• providers/aws/recommendations/parser_ri.go
• providers/aws/recommendations/parser_ri_test.go

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/cor-07-fix

---

_{Comment @coderabbitai help to get the list of available commands.}

[cristim]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

Adversarial review

The in-scope fix is correct and complete:

• parseAWSCostDetails now returns an error on present-but-unparseable UpfrontCost / EstimatedMonthlyOnDemandCost / RecurringStandardMonthlyCost, matching parseCostInformation's already-loud handling of EstimatedMonthlySavingsAmount / EstimatedMonthlySavingsPercentage.
• The caller chain (parseRecommendationDetail -> parseRecommendations) propagates with context and skips the affected detail with a warning, so a malformed CE response drops loudly instead of fabricating $0 into money math.
• Tests assert the invariant they guard (three malformed-field subtests confirmed failing pre-fix per the PR description, plus a well-formed sanity case so failures are attributable to the mutated field).
• RecurringMonthlyCost is *float64 in pkg/common.Recommendation so the nil-vs-zero distinction is preserved on the recurring side. CommitmentCost / OnDemandCost are bare float64 but with the new fail-loud contract the consumer never sees an ambiguous "0 from parse failure" anymore.

Out-of-scope siblings filed as follow-ups

The same silent-fallback shape still exists in three sibling parsers that were intentionally left out of this PR's RI scope but are equally money-affecting:

• fix(providers/aws): fail loud on unparseable SP cost fields + RI utilization hours (COR-07 siblings to #1171) #1308 -- parser_sp.go::parseOptionalFloat silently swallows parse errors for UpfrontCost / EstimatedMonthlySavingsAmount / HourlyCommitmentToPurchase / CurrentAverageHourlyOnDemandSpend on every SP recommendation, and utilization.go::parseFloat silently swallows PurchasedHours / TotalActualHours / UnusedHours (the denominator/numerator in the --target-coverage utilization signal).
• fix(providers/azure): fail loud on unparseable annualSavingsAmount / qty (COR-07 sibling to #1171) #1309 -- providers/azure/recommendations.go::extFloat / extInt silently zero annualSavingsAmount and qty; also providers/azure/services/compute/client.go::extractVMSKUCapabilities silently zeroes vCPUs / MemoryGB.

CI state (UNSTABLE) -- pre-existing, not caused by this PR

The three failing jobs all hit code this PR did not touch:

• Lint Code (errcheck) -- failures in cmd/configure_azure.go, cmd/configure_gcp.go, internal/api/handler.go, internal/auth/service_*.go, etc. Covered by PR fix(lint): clear golangci-lint debt in files untouched by open PRs #1276 (leftover lint cleanup).
• Integration Tests -- TestPostgresStore_UpsertRecommendations_AccountScopedEviction, TestPostgresStore_UpsertRecommendations_AmbientAndRegisteredCoexist, TestPostgresStore_PurchaseExecutions/Get_execution_by_ID_-_not_found in internal/config. Schema / migration concerns, not money parsing.
• Security Scanning -- JS-YAML quadratic-DoS advisory (GHSA-h67p-54hq-rp68) in NPM deps, not Go code.

These are tracked / fixable elsewhere; they do not block COR-07.

Minor (non-blocking)

• parser_ri.go:25 uses fmt.Printf("Warning: ...") while the rest of the file uses log.Printf("WARNING: ..."). Cosmetic, pre-existing — not changed here per "don't touch what you weren't asked to touch".

[cristim]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

Merge-prep: rebased onto origin/main (27fdb06). Gates: go build exit 0, go vet exit 0, gocyclo -over 10 exit 0, golangci-lint run exit 0 (root module), go test ./recommendations/... exit 0 (1.41s). Branch force-pushed; @coderabbitai review triggered.

[coderabbitai]

@cristim Thanks for the merge-prep validation. I’ll review the rebased branch.

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

Rebased onto origin/main (89658a3 -> a945344) which includes #1437 (tflint plugin cache + retry) and #1438 (trivy \*\*/.terraform skip). Workflow files confirmed to match main. Gates: go build/vet exit 0, gocyclo -over 10 exit 0 (no cycles), golangci-lint ./recommendations/... exit 0 (0 issues in touched file), go test ./recommendations/... ok. Pre-commit should be fully green.