MS Architecture Center Scanner

A maintenance and discovery tool for keeping Azure Architecture Center scenarios in sync with the Azure Pricing Calculator.

It serves two purposes:

1. Maintenance of known calculator scenarios — monitors each scenario already added to the Pricing Calculator for estimate link changes, image updates, and whether the article is still live and accessible on the Architecture Center. Flags anything that requires an update request to the calculator team.

2. Discovery of new calculator candidates — scans all Architecture Center articles to identify those that have added or updated a saved estimate link and are not yet in the calculator, so they can be evaluated for addition.

---

Context and background

What is the Azure Pricing Calculator?

The Azure Pricing Calculator is a Microsoft tool that helps customers estimate the cost of running Azure workloads. Within the calculator, there is a Scenarios section that contains pre-built cost estimates tied to specific Azure architecture patterns from the Architecture Center. Each scenario in the calculator references:

• A saved estimate link — a shared Azure Pricing Calculator estimate URL (e.g. https://azure.com/e/...) that opens a pre-configured cost estimate
• An architecture diagram image — the primary visual from the corresponding Architecture Center article
• A title — a short name for the scenario (subject to a strict character limit in the calculator UI)

What is the Azure Architecture Center?

The Azure Architecture Center is a Microsoft documentation site containing reference architectures, design patterns, and example scenarios for building solutions on Azure. It is maintained in the open-source GitHub repository MicrosoftDocs/architecture-center. Articles are regularly updated by Microsoft engineers and the community — which is why ongoing monitoring is needed.

The problem this tool solves

Architecture Center articles change over time. Authors add, update, or remove pricing estimate links. Architecture diagrams get redesigned. Entire articles get taken down or redirected. Without automated monitoring, these changes go unnoticed and the Pricing Calculator falls out of sync with the Architecture Center.

Previously, this monitoring was done manually. This tool automates it.

Who owns what

Party	Responsibility
You (PM/TPM)	Run the monthly scan, review outputs, submit update requests, maintain estimate_scenarios.xlsx
Calculator team	Receives update requests and makes changes to the Pricing Calculator scenarios
Architecture Center team	Maintains the Architecture Center articles and repo; you do not submit requests to them

Key contacts

Note for the incoming person: update this section with current contacts before handing off again.

• Calculator team contact: [add name and email]
• Architecture Center repo: https://github.com/MicrosoftDocs/architecture-center
• Your forked repo: [add your fork URL]

---

Your monthly workflow

This is the end-to-end process you follow each month. Each step is explained in detail later in this document.

Step 1 — Run the monthly scan

1. Open your forked repo on GitHub
2. Go to Actions → Architecture Scan + Estimate Comparison
3. Click Run workflow → Run workflow
4. Wait for the workflow to complete (typically 5–10 minutes)
5. Download scan-results.xlsx from the workflow artifacts

Step 2 — Review the estimate-updates tab

These are scenarios already in the Pricing Calculator whose estimate link has changed in the Architecture Center.

For each row:

1. Open the yml_url link and verify the new estimate link is correct
2. Submit an update request to the calculator team with the new estimate link
3. Update the estimate_link column in estimate_scenarios.xlsx with the new URL
4. Re-run the scan to confirm the row no longer appears

Step 3 — Review the estimate-link-removed tab

These are scenarios already in the Pricing Calculator whose estimate link has been completely removed from the Architecture Center article.

For each row:

1. Open the yml_url link to verify the estimate link is genuinely gone
2. Decide whether to retire the calculator scenario or wait to see if it is restored
3. If retiring: submit a retirement request to the calculator team
4. Update or remove the row in estimate_scenarios.xlsx

Step 4 — Review the inventory-health tab

These are availability checks on all scenarios currently in your reference inventory.

Focus on rows with inventory_status of scenario_removed or scenario_redirected:

1. Open the yml_url to verify the page is genuinely gone or redirected
2. For scenario_removed: submit a retirement request to the calculator team; remove the row from estimate_scenarios.xlsx
3. For scenario_redirected: check where the redirect leads; if the content moved to a new URL, update yml_url in estimate_scenarios.xlsx and recheck; if the content is gone, retire the scenario

Step 5 — Review the image-changes tab

These are scenarios where the primary architecture diagram has changed in the repo since the last confirmed baseline.

For each changed row:

1. Open the yml_url to review the updated diagram
2. Decide whether the change is significant enough to update the calculator
3. If updating: submit an image update request to the calculator team with the new image
4. Once the calculator team confirms the update is done, trigger the Update Image Baseline workflow (see Baseline update behaviour)

For each image_not_found row:

1. The image path stored in estimate_scenarios.xlsx is no longer valid
2. Find the new path in the Architecture Center repo and update primary_image_path in estimate_scenarios.xlsx

Step 6 — Review the new-candidates tab

These are Architecture Center articles that now have a valid saved estimate link and are not yet in the Pricing Calculator.

For each row:

1. Open the yml_url and review the article and its estimate link
2. Decide whether it is a good candidate for the calculator
3. If adding: submit an add request to the calculator team with the estimate link, title, and image
4. Add a new row to estimate_scenarios.xlsx (see Adding a new scenario to the reference inventory)

Step 7 — Commit your reference file updates

After actioning all tabs, commit and push the updated estimate_scenarios.xlsx to your forked repo so the next monthly run has an accurate baseline.

---

The reference inventory file (estimate_scenarios.xlsx)

This is the single source of truth for all scenarios you are tracking. It lives in the root of your forked repo and is read by every script in the pipeline.

Column reference

Column	Who fills it	Description
status	You	Skip to exclude a row from all processing; anything else (e.g. Published, Submitted, In Review) is processed normally
title_in_calculator	You	The scenario title as it appears in the Pricing Calculator (50 character limit enforced by the calculator UI)
yml_url	You	The canonical Architecture Center article URL (no trailing /index)
estimate_link	You	The saved estimate URL currently in the Pricing Calculator for this scenario
primary_image_path	You	Repo-relative path to the architecture diagram used in the calculator (e.g. docs/example-scenario/ai/media/diagram.svg)
primary_image_sha256	Tool	SHA-256 hash of the image file — auto-populated by the workflow. Do not edit manually.
Notes	You	Free-text notes for your own reference

Status values

The status column controls whether a row is included in processing. Only rows with status = Skip are excluded. All other values are treated as active:

Status	Meaning
Published	Scenario is live in the Pricing Calculator
Submitted	Update or add request has been submitted to the calculator team; awaiting confirmation
In Review	Under review by the calculator team
Skip	Excluded from all processing — use this for scenarios you want to keep in the file but not monitor

You can use any status value that is meaningful to your workflow — only Skip has special meaning to the tool.

Adding a new scenario to the reference inventory

When a new article is approved for the Pricing Calculator and the calculator team confirms it has been added:

1. Open estimate_scenarios.xlsx
2. Add a new row with:
   • status = your current tracking status (e.g. Published)
   • title_in_calculator = the title as entered in the calculator (≤ 50 characters)
   • yml_url = the article URL from the yml_url column in scan-results.xlsx
   • estimate_link = the saved estimate URL from the estimate_link column in scan-results.xlsx
   • primary_image_path = the repo-relative image path from the primary_image_path column in scan-results.xlsx
   • Leave primary_image_sha256 blank — the tool will populate it on the next run
3. Commit and push estimate_scenarios.xlsx to your forked repo

Updating an existing scenario

When the calculator team confirms an estimate link or image update has been applied:

1. Update estimate_link or primary_image_path in the corresponding row
2. If updating an image: also run the Update Image Baseline workflow to reset the hash (see Baseline update behaviour)
3. Commit and push estimate_scenarios.xlsx

Removing a scenario

When a scenario is retired from the Pricing Calculator:

1. Either delete the row entirely, or change status to Skip if you want to keep a record
2. Commit and push estimate_scenarios.xlsx

---

What the scanner evaluates

Two content patterns in the Architecture Center

The Architecture Center publishes articles using two distinct structures. The scanner handles both:

Pattern A — YML + MD (companion pair) A .yml file contains page metadata and a [!INCLUDE] directive pointing to a separate .md file that holds the article body. The .yml is the published page; the .md is the content source.

Pattern B — Standalone MD A single .md file contains both metadata (in a YAML front matter block at the top) and the full article body. There is no companion .yml. The .md file is itself the published page.

Both patterns appear in the Architecture Center repo. Examples of Pattern B articles include pages under networking/guide/ and data-guide/disaster-recovery/ that have no associated .yml file. The scanner's Pass 2 picks these up by finding .md files with a title field in their front matter that are not already consumed as [!INCLUDE] targets by any .yml file.

Evaluation pipeline

The scanner applies three gates in strict sequence. Each gate only runs if the previous one passed. All rows are kept in the output regardless of which gate they stopped at.

Gate	Column	Question	Stops here if…
1	scan_status	Did the file parse and resolve correctly?	File is broken or missing — scan_status = error code
2	in_scope	Is this a complete, valid scenario?	Any of the three scope criteria are missing — in_scope = FALSE
3	criteria_passed	Does the article contain a usable pricing estimate link?	No valid estimate link found — criteria_passed = FALSE

Comparison (comparison_status) only runs on rows that reach and pass Gate 3.

Gate 1 — Scan status

Before any content evaluation, the scanner checks whether each file could actually be parsed and resolved. A record receives scan_status = ok only if the file parsed successfully and all content resolved correctly. Any structural failure sets scan_status to one of the error codes below, forces in_scope = FALSE with out_of_scope_reason = scan_error, and stops all further evaluation. These rows are kept in the output for auditability.

scan_status error code	Applies to	What it means	Common cause
yaml_parse_failed	Pattern A (YML+MD)	The .yml file exists but could not be parsed as valid YAML, or its top-level value is not a dictionary	Syntax errors, encoding issues, or a non-article YAML file (e.g. a toc.yml or docfx.json picked up by the glob)
missing_content_string	Pattern A (YML+MD)	The .yml file parsed successfully but contains no content key, or its value is not a string	The YML file is a metadata-only or config file rather than an article wrapper; or the content field is structured as a list instead of a string
no_include_directive	Pattern A (YML+MD)	The content string was found but contains no [!INCLUDE] directive pointing to an .md file	The YML embeds its article body directly as inline text instead of referencing a companion .md file, which is an uncommon but valid authoring pattern the scanner does not process
include_md_unresolvable	Pattern A (YML+MD)	An [!INCLUDE] directive was found but the referenced path could not be resolved to a location within the repo	The path in the directive is malformed, uses an unsupported syntax, or points outside the docs root
include_md_missing	Pattern A (YML+MD)	The [!INCLUDE] path resolved correctly but no file exists at that location on disk	The companion .md file was deleted, renamed, or never committed; or the YML references a file in a different branch

Gate 2 — Scope filter (in / out of scope)

Before any pass/fail evaluation runs, the scanner applies a scope filter. A scenario is in scope (in_scope = TRUE) only when all three of the following are present:

1. Non-blank title — pulled from the YML metadata or MD front matter
2. Non-blank description — same source
3. At least one architecture image — at least one image reference (:::image, ![](), <img>, etc.) must appear anywhere in the article body, in any format and whether local or externally hosted

azureCategories is captured in the output for every scenario but a missing or blank category does not cause a scenario to be marked out of scope.

Scenarios that fail one or more criteria receive in_scope = FALSE and an out_of_scope_reason listing each failing criterion (semicolon-separated): blank_title, blank_description, or no_architecture_image. All rows are preserved in the output — out-of-scope rows are visible in scan-results for auditability but are excluded from pass/fail evaluation, comparison, and the action queue tabs.

Gate 3 — Primary question (pass / fail)

The scanner answers the primary question: Does the Architecture Center article include a usable pricing estimate link?

A scenario passes (criteria_passed = TRUE) if the included Markdown article contains at least one of the following:

1. Azure Experience links — https://azure.com/e/*
2. Pricing Calculator shared estimate links — https://azure.microsoft.com/pricing/calculator?...shared-estimate=*

Important: Pricing Calculator tool/root links (for example, /pricing/calculator without a saved estimate) do not count as usable estimates.

Failure reasons

If no usable estimate is found, the scenario fails with one of the following reasons:

• no_estimate_link_calculator_tool_link_only — Pricing Calculator links exist, but they are tool/root links only (no saved estimate).
• no_estimate_link — No Pricing Calculator links of any kind were found in the article.

Images (informational only)

Images are detected for every included .md article and captured in the output (image_download_urls, primary_image_path). Image presence does not affect pass/fail and is provided for context and review only.

---

Estimate comparison (post-scan)

For scenarios where criteria_passed = TRUE, the tool compares detected usable estimate links against the reference inventory in estimate_scenarios.xlsx.

Comparison behavior

• Scenarios are matched using the Learn URL (yml_url).
• Estimate URLs are normalized before comparison (case, whitespace, and trailing slashes removed; only the identity-defining shared-estimate query parameter is retained).
• If an article contains multiple valid estimate links (for example, Small / Medium / Large cost profiles), all compliant links are preserved in the report (newline-separated).
• A scenario is treated as a match if any scanned estimate link matches the inventory estimate link after normalization.
• Only rows in estimate_scenarios.xlsx with status != Skip participate in comparison.

comparison_status values

comparison_status value	Meaning	Action queue
matched_existing_scenario_same_estimate	Scenario exists in inventory and the scanned estimate link matches	No action needed
matched_existing_scenario_new_estimate	Scenario exists in inventory but the estimate link has changed to a different URL	estimate-updates tab
estimate_link_removed	Scenario exists in inventory, article is still live and in scope, but the estimate link has been removed from the article entirely	estimate-link-removed tab
new_estimate_candidate	Scenario has a valid estimate link but does not exist in the inventory	new-candidates tab
not_applicable	Scenario failed Gate 2 or Gate 3; comparison not performed	—

Only scenarios with in_scope = TRUE and criteria_passed = TRUE participate in estimate comparison.

Action queue worksheets

The Excel output includes three dedicated action queue worksheets:

estimate-updates — scenarios already in the inventory whose estimate link has changed to a different URL. Each row requires submitting an update request to the calculator team and updating estimate_scenarios.xlsx.

estimate-link-removed — scenarios already in the inventory whose estimate link has been removed from the Architecture Center article entirely while the page is still live. Each row requires submitting a retirement request to the calculator team and updating estimate_scenarios.xlsx.

new-candidates — articles not yet in the inventory that now have a valid saved estimate link. Each row is a candidate for evaluation and potential addition to the calculator. Includes primary_image_path to make it easy to populate estimate_scenarios.xlsx when adding a new row.

---

Inventory health check

After the estimate comparison runs, the workflow performs a dedicated health check on every non-Skip scenario in estimate_scenarios.xlsx. This detects scenarios that have been taken down or redirected on the Architecture Center so they can be retired from the Pricing Calculator.

How it works

The check runs in two stages for each inventory scenario:

Stage 1 — Reverse lookup (no network calls) Checks whether the scenario's yml_url appears anywhere in the current scan-results.xlsx. If it does not, the source file has been deleted or renamed in the repo and the scenario is immediately flagged as scenario_removed without making any HTTP request.

Stage 2 — HTTP liveness check (network) For scenarios whose file still exists in the repo, makes an HTTP GET request to the live yml_url on learn.microsoft.com and follows any redirects. This catches three situations the reverse lookup cannot: a page removed from publishing while its source file still exists in the repo, a page that redirects to a different URL, and a page that returns a 404.

Requests are throttled at one per second by default to be polite to learn.microsoft.com. The throttle interval can be adjusted with the --throttle flag.

inventory_status values

Value	Stage detected	Meaning	Recommended action
active	Stage 2	URL resolves correctly to the expected page	No action needed
scenario_removed	Stage 1 or 2	Source file deleted from repo, or URL returns 404	Retire or redirect this Pricing Calculator scenario
scenario_redirected	Stage 2	URL permanently redirects to a different page	Review the redirect target; update or retire the calculator scenario
scan_error	Stage 1	File exists but failed scanner Gate 1	Fix the source file structural issue first, then re-run
out_of_scope	Stage 1	File exists and parses but is currently out of scope	Investigate why the scenario dropped out of scope

inventory-health worksheet columns

• yml_url — The canonical URL from estimate_scenarios.xlsx
• title_in_ac — Article title from the Architecture Center (scanner output)
• title_in_calculator — Title as it appears in the Pricing Calculator (from estimate_scenarios.xlsx)
• estimate_link — Estimate link from estimate_scenarios.xlsx
• inventory_status — One of the five status values above
• redirect_target — The final URL after following redirects (blank if no redirect)
• http_status_code — HTTP response code from the liveness check (blank for scenario_removed via Stage 1)
• note — Human-readable explanation of the status

---

Image change detection

After the inventory health check, the workflow checks whether the primary architecture diagram for each non-Skip scenario in estimate_scenarios.xlsx (that has a primary_image_path) has changed since the last confirmed baseline.

How it works

For each eligible row in estimate_scenarios.xlsx, the script:

1. Resolves the image file on disk (the Actions runner has a full repo clone, so no network calls are needed)
2. Computes a SHA-256 hash of the file's raw bytes
3. Compares that hash against the primary_image_sha256 column in estimate_scenarios.xlsx
4. Records the result in the image-changes tab of scan-results.xlsx

SHA-256 hashing detects any change to the image content — including cases where a contributor replaces an image file in-place without changing the filename.

image_change_status values

Value	Meaning	Action
unchanged	Hash matches the stored baseline	No action needed
changed	Hash differs from baseline — image was updated in the repo	Review the updated diagram; submit an image update request to the calculator team if needed
new_baseline	No stored hash yet (first run or new inventory row)	Hash recorded as the new baseline; no comparison was made this run
image_not_found	primary_image_path does not exist on disk	The file may have been moved, renamed, or deleted; verify and update the path in estimate_scenarios.xlsx
skipped	Row has status = Skip or blank primary_image_path	No action needed

image-changes worksheet columns

• yml_url — Architecture Center article URL
• title_in_ac — Article title from the Architecture Center
• title_in_calculator — Title as it appears in the Pricing Calculator
• primary_image_path — Repo-relative path to the tracked image
• image_change_status — changed or image_not_found
• stored_sha256 — The hash from the previous baseline (blank for image_not_found)
• current_sha256 — The hash computed this run (blank for image_not_found)
• note — Human-readable explanation

Reference file image columns

• primary_image_path — Repo-relative path to the image used in the Pricing Calculator (e.g. docs/example-scenario/ai/media/diagram.svg). You fill this in manually when adding a new row.
• primary_image_sha256 — SHA-256 hash of the image file, auto-populated by the workflow. Do not edit this column manually.

Baseline update behaviour

The image hash baseline is not updated automatically during the monthly scan. This is intentional — a detected change must stay visible across every run until you have confirmed the update and actioned it.

The workflow for updating the baseline is:

1. Monthly scan runs → image-changes tab flags a scenario with changed
2. You review the updated image and decide it needs to be updated in the calculator
3. You submit an image update request to the calculator team
4. Calculator team updates the scenario with the new image and confirms
5. You manually trigger the Update Image Baseline workflow from GitHub Actions
6. The workflow recomputes all hashes, saves them to estimate_scenarios.xlsx, and commits the file back to the repo
7. The next monthly scan compares against this newly confirmed baseline

Important: do not run the Update Image Baseline workflow to dismiss a detected change without first updating the Pricing Calculator. Doing so will silently lose the signal and the change will not be detected again unless the image changes further.

---

Repository files and what they do

File	Purpose
scripts/scan_architecture_center_yml.py	Scans all Architecture Center YML and MD files in the repo. Handles both Pattern A (YML+MD) and Pattern B (standalone MD). Produces scan-results.json.
scripts/build_scan_results_xlsx.py	Converts scan-results.json into scan-results.xlsx with the scan-results worksheet.
scripts/run_compare_only.py	Compares scanned articles against estimate_scenarios.xlsx. Produces estimate-updates, estimate-link-removed, new-candidates, and summary worksheets.
scripts/run_inventory_health.py	Runs reverse lookup and HTTP liveness checks on all non-Skip inventory scenarios. Produces inventory-health worksheet and appends to summary.
scripts/run_image_check.py	Hashes primary images and compares against stored baselines (detect mode). Or resets baselines after changes are actioned (update-baseline mode). Produces image-changes worksheet and appends to summary.
estimate_scenarios.xlsx	Your reference inventory. The single source of truth for all tracked scenarios. Read by run_compare_only.py, run_inventory_health.py, and run_image_check.py. Updated by you after actioning changes; primary_image_sha256 is auto-updated by the workflow.
.github/workflows/scan_and_compare.yml	Monthly scan workflow. Runs all five steps in sequence and uploads scan-results.xlsx as an artifact.
.github/workflows/update_image_baseline.yml	Manual baseline reset workflow. Run only after confirming the Pricing Calculator has been updated with the new image.

---

How to get started (first-time setup)

Prerequisites

• Access to your team's forked copy of MicrosoftDocs/architecture-center on GitHub
• Write access to the repo (needed to commit estimate_scenarios.xlsx updates)
• estimate_scenarios.xlsx already populated with the current inventory of tracked scenarios

If the repo fork does not exist yet:

1. Fork https://github.com/MicrosoftDocs/architecture-center
2. Copy the scanner files into the forked repo at the paths shown in the table above
3. Add estimate_scenarios.xlsx to the root of the forked repo

Running the scan for the first time

1. Open the forked repo on GitHub
2. Go to Actions → Architecture Scan + Estimate Comparison
3. Click Run workflow → Run workflow
4. Wait for the workflow to complete

What to expect on the first run:

• All rows in estimate_scenarios.xlsx that have a primary_image_path but no primary_image_sha256 will show new_baseline in the image check — this is expected. The hashes are recorded and the image-changes tab will be empty.
• The Update Image Baseline workflow does not need to be run on the first scan — hashes are recorded automatically when a row has no existing baseline.
• All existing inventory scenarios should appear as matched_existing_scenario_same_estimate or active if the reference file is up to date.

Downloading results

After the workflow completes:

1. Click the completed workflow run
2. Scroll to Artifacts at the bottom
3. Download scan-results — it contains scan-results.xlsx

---

Outputs and how to interpret them

After a successful run, scan-results.xlsx contains the following worksheets:

Worksheet	Contents	When to use it
scan-results	All scanned articles (~500+ rows)	Full audit view; look here when investigating a specific article
estimate-updates	Inventory scenarios with a changed estimate link	Monthly action: submit update requests
estimate-link-removed	Inventory scenarios whose estimate link was removed	Monthly action: submit retirement requests
new-candidates	Articles not in inventory with a valid estimate link	Monthly action: evaluate for addition to calculator
inventory-health	Availability check for all inventory scenarios	Monthly action: flag removed or redirected pages
image-changes	Inventory scenarios with a changed or missing image	Monthly action: submit image update requests
summary	Counts for every metric across all steps	Quick health check; review first to understand the scale of changes

Key columns in scan-results

• title_in_ac — Article title from the Architecture Center
• description — Article description
• azureCategories — Azure solution categories
• ms.date — Article freshness date; useful for prioritization
• yml_url — Published Architecture Center article URL
• image_download_urls — All images found in the article (informational)
• primary_image_path — Repo-relative path to the first image found; use this to populate estimate_scenarios.xlsx when adding a new row
• estimate_link — One or more usable estimate links (newline-separated)
• scan_status — Gate 1 result: ok or a structural error code (see Gate 1)
• in_scope — Gate 2 result: TRUE or FALSE
• out_of_scope_reason — Why the scenario is out of scope (blank_title, blank_description, no_architecture_image, or scan_error)
• criteria_passed — Gate 3 result: TRUE or FALSE
• failure_reason — Why criteria_passed is FALSE (no_estimate_link or no_estimate_link_calculator_tool_link_only)
• comparison_status — Estimate comparison result; drives the action queue tabs
• yml_path — Path to the source YML file in the repo
• include_md_path — Path to the included MD file (Pattern A) or same as yml_path (Pattern B)
• md_author_name / md_ms_author_name — Article author; useful for ownership context

How to interpret the summary tab

The summary tab is the first thing to check after downloading results. It gives you a count for every metric across all five pipeline steps:

• Gate 1/2/3 counts — how many articles passed or failed each gate
• Comparison status counts — how many matched, updated, removed, or are new candidates
• Action queue counts — rows in each action queue tab
• Inventory health counts — how many scenarios are active, removed, or redirected
• Image change counts — how many images are unchanged, changed, or missing

If all counts look normal (most scenarios matched, few or no changes), a quick scan of the action queue tabs is all you need. If counts look unexpected, use the scan-results tab to investigate.

---

Troubleshooting

A scenario I expect to see is missing from scan-results

Check 1: Is the article's source file a Pattern A (index.yml) file? The scanner strips /index from the URL — so the yml_url in scan-results.xlsx will be .../folder-name, not .../folder-name/index. If your reference file still has /index in the URL, update it.

Check 2: Does the article have a title, description, and at least one image? If any are missing, it will be in scan-results but with in_scope = FALSE. Filter the scan-results tab on in_scope = FALSE and check out_of_scope_reason.

Check 3: Did the article recently change its structure? A Gate 1 error (scan_status != ok) means the scanner found the file but couldn't process it. Check failure_reason for the specific error code.

A scenario appears as new_estimate_candidate but it's already in my reference file

The yml_url in estimate_scenarios.xlsx likely doesn't match what the scanner produced. Common causes:

• The reference file has /index at the end of the URL but the scanner strips it (see above)
• A trailing slash difference — normalize all URLs to no trailing slash

The inventory health check flagged a scenario as scenario_redirected but the article seems fine

Some Architecture Center articles use /index in their URL at the source but the docs platform normalizes it — this can appear as a redirect. Check if the redirect target is just the same URL without /index. If so, update the yml_url in your reference file to the canonical form (without /index).

The image baseline check shows changed but I didn't request any image updates

Architecture Center authors regularly refresh diagrams. A changed result means the image file bytes are different from when you last confirmed the baseline — it doesn't necessarily mean the change is significant. Open the article and review the diagram before deciding whether to submit an update request.

The workflow failed mid-run

Check the Actions log for the specific step that failed. Common causes:

• Python syntax error in a script (check recent code changes)
• estimate_scenarios.xlsx missing a required column
• Network timeout during the inventory health HTTP checks (re-run the workflow)