Conversation
Rename the tags: frontmatter to keywords: across all Cloud-to-Cloud source docs, preserving the existing values. GA4 shows the generated /help/docs/tags/* pages get ~99.7% bot traffic and 17 organic views total over 90 days, no SEO/GEO value, so the tags feature itself is being retired sitewide. The terms are still useful for search, so they move into keywords: instead of being dropped. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Drop the unused tags: [cloud soar integrations] frontmatter. Unlike the C2C source docs, this page had no pre-existing keywords: field and the single tag value isn't worth introducing one for, so it's just removed. Also includes an incidental trailing-whitespace fix on the Password bullet. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Remove tags: from the style guide's example frontmatter and the two scaffold templates (app-template-v2, c2c-source), renaming it to keywords: in the templates so new docs keep the same search terms without generating a tag page. Mark tags: as deprecated in the style guide's frontmatter parameter table so it isn't reintroduced. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Remove tags: from the docusaurus and sumo-style skill frontmatter examples, and rename tags to keywords in app-doc.md's required substitutions, checklist, and worked example so the /app-doc command matches the updated app-template-v2.md scaffold. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…p config Drop the Disallow: /tags/* and Disallow: /*?tags= rules from robots.txt, and /tags/** from the sitemap ignorePatterns, now that no tag routes exist to disallow or ignore. Note: this repo's robots.txt is served at /help/robots.txt, not the production domain root (https://www.sumologic.com/robots.txt, a separate WordPress-managed file that never included these rules), so this was cleanup rather than a fix to something that was blocking crawlers in production. The Algolia Crawler Admin's exclusionPatterns (external, not in this repo) also has two tags-related entries that are now no-ops; flagged separately for manual removal there. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The prior redirects commit covered every tag that mapped to exactly one doc, but missed the 3 tags that were shared across multiple docs: cloud-to-cloud (82 docs), cloud-siem-enterprise (4 docs), and cyberark (2 docs, the live one plus an unlisted beta variant). Verified using Docusaurus's actual lodash.kebabCase tag-slug algorithm against every distinct tag value that existed on main, pre-PR: 96 total, now all 96 have a redirect (was 93/96). Shared tags redirect to the C2C source list index page, since no single vendor doc is a fair target for a page that covered several. cyberark redirects to the live doc, not the unlisted beta one. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
amee-sumo
left a comment
amee-sumo
left a comment
There was a problem hiding this comment.
Left inline comments on 6 broken redirect targets in cid-redirects.json. Full context in my top-level review comment. Everything else I checked (the tags->keywords renames, zero remaining tags: frontmatter repo-wide, robots.txt/sitemap cleanup, style guide/template/tooling updates) is clean.
Fix redirects that assumed every C2C doc's filename ends in -source, which isn't true for bitwarden.md, crowdstrike-fdr-host-inventory.md, crowdstrike-filevantage.md, and jfrog-xray.md, plus one hyphenation mismatch (intel-471-threat-intel-source.md, not intel471-...). All 6 were anyone hitting the old tag URL landing on a 404 instead of the real doc. Found by amee-sumo in PR review. Verified by checking every one of the 97 /docs/tags/* redirect targets against the actual filesystem, not just the 6 flagged — no other mismatches exist. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
Fixed all 6 broken redirect targets in f5d17e2. Replied on each thread with what it now points to. While fixing them I checked every one of the 97 /docs/tags/* redirects against the actual filesystem, not just the 6 you flagged. No other mismatches. Also re-ran the earlier slug-completeness check (every distinct tag value that ever existed on main, matched against Docusaurus's real lodash.kebabCase algorithm) and it's still at 0 missing. |
My review comments on this PR flagged these 5 as broken by checking their target against the filesystem filename. That was the wrong check: this repo's docs derive their actual URL from the `id:` frontmatter field, not the filename, and sidebars.ts confirms these routes are id-based. The original targets (before my requested fix) were already correct; the fix commit broke them by matching the filename instead. The build failure this caused is what surfaced the mistake. Reverts: - bitwarden -> bitwarden-source - crowdstrike-fdr-host-inventory -> crowdstrike-fdr-host-inventory-source - crowdstrike-filevantage -> crowdstrike-filevantage-source - intel-471-threat-intel-source -> intel471-threat-intel-source - jfrog-xray -> jfrog-xray-source (both /docs/tags/jfrog and /docs/tags/xray) Verified all 97 /docs/tags/* redirects by building a route map from every doc's actual id: frontmatter (or slug:, or category-index behavior), not just filesystem paths. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
I owe a correction on my earlier review comments. I flagged 6 redirect targets as broken by checking them against the filesystem filename, but that was the wrong check — this repo derives each doc's actual URL from its I've pushed a commit reverting those 5 entries back to their original, correct values, and re-verified all 97 Sorry for the back-and-forth this caused — the PR itself was correct on this point from the start. |
amee-sumo
left a comment
amee-sumo
left a comment
There was a problem hiding this comment.
run-build-checks now passes. All 97 /docs/tags/* redirects are verified against the real routing (id: frontmatter, not filenames), the tags->keywords renames are clean, zero tags: frontmatter remains repo-wide, and robots.txt/sitemap/style-guide/tooling updates all check out.
Approving.
2 participants
Purpose of this pull request
This pull request removes the unused Docusaurus tag taxonomy sitewide. GA4 shows the ~120 auto-generated
/help/docs/tags/*pages get almost no organic traffic and no established reader benefit, while costing ongoing upkeep (drift againstkeywords:, another thin crawl-surface URL per tag). See DOCS-1879 for the full rationale, including verified GA4 numbers and technical findings.Five commits, one per batch:
docs/send-data/hosted-collectors/cloud-to-cloud-integration-framework/):tags:renamed tokeywords:, values unchanged.zip-tools.md:tags:removed (no existingkeywords:field, single value not worth preserving).tags:guidance removed/deprecated so contributors don't reintroduce it; templates renamed tokeywords:to keep new docs consistent with (1)./app-doccommand): same treatment, so the doc-creation slash commands match the updated templates.robots.txt+ sitemap config: removed the now-obsoleteDisallow: /tags/*/Disallow: /*?tags=rules and the/tags/**sitemap-ignore entry.Verified with a clean local
yarn build: zerodocs-tags-*route artifacts generated, confirming the docs plugin stops generating tag routes once no doc hastags:frontmatter (there's no config flag for this — route generation is driven purely by frontmatter presence).Not in this repo, flagged separately: the Algolia Crawler Admin's
exclusionPatternshas two now-obsolete tag-related entries (harmless no-ops, not blocking) to remove manually in the Algolia dashboard.Select the type of change
Ticket (if applicable)
https://sumologic.atlassian.net/browse/DOCS-1879
🤖 Generated with Claude Code