You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Expose typed detector outcomes and inspectable capacity evidence. Correct sizing, fan-out, storage and configuration predicates, distinguish structural merge findings from action eligibility, and evaluate authoritative shared offers without duplicating recoverable capacity.
Includes backup/change-feed metadata, collector wiring, resource-oriented UI wording, native regressions and generated localization. Carries the existing raw-log observation duration into the stricter detector contract; aggregate-log acquisition follows in the next PR.
Stack
Part 2 of 9 in the Account Overview advisory stack. Depends on #3566. Base: dev/dshilov/account-overview-01-offer-metrics.
Each subsequent PR targets its predecessor branch so GitHub can represent the chain as a native PR stack.
Validation
The integrated stack passed localization generation, oxfmt, lint, all TypeScript build targets, and all unit suites. The full unit command used VITEST_MAX_WORKERS=2 to avoid local resource-contention timeouts. Intermediate runtime branches also passed builds and focused regressions.
Review fixes are committed to their originating PR layers; descendants are rebased onto the updated predecessors.
Tip: the HTML report artifact contains a self-contained Playwright report.
Download the zip, extract, and open index.html — or run npx playwright show-report <extracted-dir> for the interactive view.
The reason will be displayed to describe this comment to others. Learn more.
🟡 Changes recommended
The 200-shape cap can undercount fan-out execution rates and suppress both detectors, requiring uncapped classification aggregates and regression coverage for workloads exceeding 200 shapes.
Part 2 of the Account Overview advisory stack builds on #3566’s offer-scoped metrics to refine native detector decisions and expose their supporting evidence.
Changes:
Adds typed evaluation outcomes, capacity evidence, and separate action-eligibility decisions.
Updates sizing, fan-out, storage, and configuration rules, including shared-offer ownership and backup/change-feed metadata.
Adds regression coverage and resource-oriented UI wording.
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
The documented Coda commands and 379-scenario result are not reproducible because this layer lacks the required package.json scripts and configuration.
Document only Coda checks implemented in this stack layer
specs/account-overview.md:684
This stack layer's package.json defines neither externalSpecifications.coda nor the test:coda and coda:check scripts, so contributors cannot run the documented commands or reproduce the claimed 379-scenario result. Describe the native regressions available here and defer the harness, pin and workflow guarantees until their implementation lands. The same issue affects the freshness-check paragraph at specs/account-overview.md:709–713.
The reason will be displayed to describe this comment to others. Learn more.
🟡 Changes recommended
The documentation claims Coda verification and conformance results that cannot be reproduced with this revision’s tooling, so those claims should be deferred or labeled as planned.
The reason will be displayed to describe this comment to others. Learn more.
🟡 Changes recommended
A valid zero-byte observation is incorrectly classified as insufficient data rather than below the materiality floor, and the behavior needs correction with regression coverage.
The reason will be displayed to describe this comment to others. Learn more.
🟡 Changes recommended
Regression coverage is missing for the new regional-capacity calculations, including mixed shared and dedicated manual offers, non-production versus legitimate-HA removable capacity, and autoscale or incomplete inventory exclusions.
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
Shared-offer sizing depends on Azure Monitor’s empty-CollectionName scoping, which remains unverified against populated mixed shared/dedicated databases and cannot be established by SDK fixtures.
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
Human live-service verification is needed because shared-offer sizing and saturation depend on Azure Monitor’s empty CollectionName scope, which remains unvalidated on a populated mixed shared/dedicated database.
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
Live validation of empty-CollectionName throughput metrics and dedicated-traffic exclusion remains outstanding for a populated mixed database, and controlled fixtures cannot establish Azure metric semantics.
No open issues 🎉
Review effort: Balanced
Dmitrii Shilov (bk201-)
changed the title
[2/8] Align native Account Overview advisory evaluations
[2/9] Align native Account Overview advisory evaluations
Nov 18, 2026
Splitting out configuration failures left filter/map as the only array member in the native-detector and request-log layers, triggering unicorn/no-useless-spread.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The top-200 shape cap biased execution rates and shares, suppressing findings or overstating structural misalignment.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Keep PR two accurate on its own; restore the pin, commands and scenario evidence with PR six.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The zero-initialized maximum incorrectly treated a valid zero-byte sample as missing evidence. Use the empty partition list and missing-history flag for insufficient_data, preserving too_small for observed zero bytes.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Exercise a mixed shared/dedicated all-manual inventory through the real advisory engine, excluding autoscale and incomplete inventories. Verify removable non-production capacity separately from legitimate production HA.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Describe only the raw request-log acquisition available in this PR; aggregate selection belongs to the following acquisition layer.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The reason will be displayed to describe this comment to others. Learn more.
🟡 Changes recommended
Missing metrics are treated as zero, and partition series may be truncated, risking false decommission guidance and undersized throughput recommendations.
Remove the obsolete container-count key at the earliest affected PR so each descendant passes l10n:check after rebasing. Regenerated the catalog with npm run l10n.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The reason will be displayed to describe this comment to others. Learn more.
🔵 Needs a closer look
The changes can misreport throttling, incorrectly flag supported MongoDB configurations, and delay advisories through excessive per-container metric requests.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Expose typed detector outcomes and inspectable capacity evidence. Correct sizing, fan-out, storage and configuration predicates, distinguish structural merge findings from action eligibility, and evaluate authoritative shared offers without duplicating recoverable capacity.
Includes backup/change-feed metadata, collector wiring, resource-oriented UI wording, native regressions and generated localization. Carries the existing raw-log observation duration into the stricter detector contract; aggregate-log acquisition follows in the next PR.
Stack
Part 2 of 9 in the Account Overview advisory stack. Depends on #3566. Base: dev/dshilov/account-overview-01-offer-metrics.
Each subsequent PR targets its predecessor branch so GitHub can represent the chain as a native PR stack.
Validation
The integrated stack passed localization generation, oxfmt, lint, all TypeScript build targets, and all unit suites. The full unit command used VITEST_MAX_WORKERS=2 to avoid local resource-contention timeouts. Intermediate runtime branches also passed builds and focused regressions.
Review fixes are committed to their originating PR layers; descendants are rebased onto the updated predecessors.