Fix broken external links found with automated link checking

[Copilot]

Addresses broken external URLs across the documentation.

Broken link fixes (~150 files)

• Broadcom TechDocs migrations: docs.pivotal.io and docs.vmware.com links updated to current techdocs.broadcom.com paths for Spring Cloud Services → Application Services, SSO, CredHub, MySQL, Postgres, RabbitMQ, Redis/Valkey, GemFire Native Client, Metric Registrar, Volume Services, CSB (AWS/Azure/GCP), Apps Manager
• Microsoft: docs.microsoft.com → learn.microsoft.com, docs.asp.net → learn.microsoft.com, visualstudio.com → visualstudio.microsoft.com; reduced redirects where possible without forcing localization
• Spring: projects.spring.io/spring-cloud → spring.io/projects/spring-cloud-*; trailing slash fixes
• GitHub repo moves: spring-cloud-samples/hystrix-dashboard → Netflix-Skunkworks/hystrix-dashboard, cf-platform-eng/mssql-server-broker → cloudfoundry-attic/mssql-server-broker, cloudfoundry-incubator → cloudfoundry
• CredHub API: credhub-api.cfapps.io → docs.cloudfoundry.org/api/credhub/version/main/ with corrected anchors (#_overwriting_credential_values, #_permissions)
• Discontinued services: Hystrix Dashboard marked as deprecated (no longer available), Wavefront trial references removed, dead pivotal.io/platform-acceleration-lab links removed
• Naming: "PCF" → "Tanzu Platform", "TAS Apps Manager" → "Apps Manager", "TAS SSO" / "TAS Single-Sign-on" → "Single Sign-On", "Azure Open Service Broker" → "Open Service Broker for Azure"
• Misc: RabbitMQ docs path updates, tanzu.vmware.com/gemfire → vmware.com/products/..., tanzu.vmware.com/observability → broadcom.com/products/..., hhttps typo, start.steeltoe.io/api/config/ → /api/uiconfig/

Also fixes root-relative markdown links that docfx does not see (inadvertently broken in #363)

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Scan for broken links to external URLs</issue_title>
<issue_description>We should run a tool that recursively follows all the links and reports which ones are broken. And then correct or remove them.

For example, there are many links pointing to docs.pivotal.io, such as https://docs.pivotal.io/spring-cloud-services/1-5/common/circuit-breaker/, which today redirects to the generic Broadcom page at https://techdocs.broadcom.com/us/en/vmware-tanzu.html.</issue_description>

Comments on the Issue (you are @copilot in this section)

@bart-vmware Remaining _hard_ broken links after the fixes in #385 are listed in the table below. Note this does not include soft broken links, such as the one mentioned above.

URL	Anchor Text	Linked From
https://credhub-api.cfapps.io/	CredHub API	[https...</comment_new>

• Fixes Scan for broken links to external URLs #375

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[TimHess]

@copilot this PR will merge after #426. Change the base branch and re-evaluate your changes, as some of the changes (eg: "Converted 30+ internal documentation links from .md to .html extension") do not align with how docfx works.

Additionally, when the workflow is run in the context of a PR, include a table of newly-broken links as a single comment on that PR that is updated on subsequent runs. Issues should only be created if the context was NOT a PR

[Copilot]

@copilot this PR will merge after #426. Change the base branch and re-evaluate your changes, as some of the changes (eg: "Converted 30+ internal documentation links from .md to .html e...

Addressed in cf944da:

• Reverted all .md→.html internal link changes — docfx uses .md extensions per the conventions in Set up agent instructions #426.
• Workflow updated: In PR context, broken links are posted as a single PR comment (updated on subsequent runs via message-id). Issues are only created outside PR context, and existing open issues with the same title are updated instead of duplicated.
• External link fixes (CredHub API, Spring AMQP, Spring Cloud, Samples) are retained.

I can't change the base branch myself, but I see you've already done that.

[Copilot]

Just as a heads up, I was blocked by some firewall rules while working on your feedback. Expand below for details.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• docs.cloudfoundry.org
  • Triggering command: /home/REDACTED/work/_temp/ghcca-node/node/bin/node /home/REDACTED/work/_temp/ghcca-node/node/bin/node --enable-source-maps /home/REDACTED/work/_temp/copilot-developer-action-main/dist/index.js (dns block)
• techdocs.broadcom.com
  • Triggering command: /home/REDACTED/work/_temp/ghcca-node/node/bin/node /home/REDACTED/work/_temp/ghcca-node/node/bin/node --enable-source-maps /home/REDACTED/work/_temp/copilot-developer-action-main/dist/index.js (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[TimHess]

@copilot update the check-links workflow to filter out links from the files in /docs/api. Read the output from the workflow in #424 (comment) and fix as many of those redirects as you can. Also check if there's an easy way to handle rate limiting and implement it if there is.

[Copilot]

@copilot update the check-links workflow to filter out links from the files in /docs/api. Read the output from the workflow in #424 (comment)...

Done in 04a608b:

• Workflow: Added --exclude-path 'docs/api' to skip generated API docs. Added --max-retries 6 and --retry-wait-time 30 for rate limiting.
• Fixed redirects across 75 files:
  • docs.microsoft.com → learn.microsoft.com (61 files)
  • docs.asp.net → learn.microsoft.com (10 files)
  • tanzu.vmware.com/content/blog → blogs.vmware.com/tanzu (5 files)
  • docs.pivotal.io/spring-cloud-services/ → Broadcom tech docs (11 occurrences)
  • Misc: dotnetfoundation.org, devblogs, cncf.io, techcommunity
• Fixed hard 404s: kubernetes.io services path, thepollyproject.org → Polly GitHub, loggregator develop → main branch
• Removed dead pivotal.io/platform-acceleration-lab links (expired SSL, service discontinued)

[TimHess]

@copilot

• workflow updates
  • remove opening/managing an issue, lets stick with the output being a comment when run during PR or just job output when run manually
  • remove the schedule trigger
  • remove start.steeltoe.io from the ignore
  • also ignore: fortuneservice, *.internal, and consul-register-example
  • set host-concurrency to 5 and retry-wait-time to 15
• broadcom techdocs links:
  • spring cloud services is now application services, links should look more like https://techdocs.broadcom.com/us/en/vmware-tanzu/spring/application-services/3-3/appservices/
  • Tanzu MySQL is now here https://techdocs.broadcom.com/us/en/vmware-tanzu/platform/tanzu-mysql-tanzu-platform/3-3/mysql-tp/index.html
  • csb for gcp is now here https://techdocs.broadcom.com/us/en/vmware-tanzu/platform/cloud-service-broker-gcp/1-6/csb-gcp/installing-with-gcp.html
  • cloud foundry platform links may need to point to vmware-tanzu/platform/tanzu-platform/10-3/, a complete example is https://techdocs.broadcom.com/us/en/vmware-tanzu/platform/tanzu-platform/10-3/about-tnz-platform/index.html
• gemfire native docs can now be found over here: https://techdocs.broadcom.com/us/en/vmware-tanzu/data-solutions/tanzu-gemfire-native-client/10-4/gf-nc/dotnet-about-client-users-guide.html
• for broken Steeltoe samples links:
  • with 2.x samples, check if adding AspDotNet4 or AspDotNetCore fixes them
  • anything with AspDotNet4 now only exists in the 2.x branch
  • either remove the links to the 3.x version of Freddy's BBQ or switch them to use the last commit before Drop Freddys BBQ Samples#365
• learn.microsoft.com links
  • some may be fixed by dropping .html from the end
  • never include localization (eg: /us/en), even if a redirect occurs (it is noise, but the noise would be better than forcing bad localization... can we filter those out of the report without losing anything? we'll revisit that as the last step if the only way is completely ignoring redirects, share your findings)
• trailing slash redirects are trivial, but they add noise in the report, let's remove as much noise as we can
• broken links in docs/docs/v2 may have been fixed in /v3 or /v4, especially for any Pivotal links - cross-reference equivalent pages for specific updates

[Copilot]

@copilot

• workflow updates
  • remove opening/managing an issue, lets stick with the output being a comment when run during PR or just job output when run manually
    ...

Done in 1b19072. Summary of changes:

Workflow:

• Removed schedule trigger, issue creation/management
• Removed start.steeltoe.io from ignore
• Added excludes for fortuneservice, *.internal, consul-register-example
• Set host-requests to 5, retry-wait-time to 15

Broadcom links:

• Spring Cloud Services → vmware-tanzu/spring/application-services/3-3/appservices/
• Tanzu MySQL → tanzu-mysql-tanzu-platform/3-3/mysql-tp/index.html
• CSB GCP → cloud-service-broker-gcp/1-6/csb-gcp/installing-with-gcp.html
• Cloud Foundry platform → tanzu-platform/10-3/about-tnz-platform/index.html
• GemFire → tanzu-gemfire-native-client/10-4/gf-nc/dotnet-about-client-users-guide.html
• All docs.pivotal.io (61 links) and docs.vmware.com links mapped to broadcom equivalents

Samples:

• dev→2.x for AspDotNet4 links, fixed RedisDataProtectionKeyStore path
• Removed FreddysBBQ links/references from v3 docs

Microsoft links:

• Removed /en-us/ localization, dropped .html suffixes
• Fixed trailing slash redirects (spring.io)

Regarding redirect filtering: The learn.microsoft.com locale redirects (/aspnet/ → /en-us/aspnet/) are server-side auto-redirects. Lychee has no option to ignore only locale redirects without ignoring all redirects. We could add --exclude 'learn.microsoft.com' to skip them entirely, but then we'd miss actual broken Microsoft links. Since we already removed /en-us/ from all our links, the remaining noise is just the server adding it back. Happy to revisit if you'd like to try a different approach.

[github-actions]

External link checking results

Status	Count
🔍 Total	1880
✅ Successful	927
⏳ Timeouts	0
🔀 Redirected	143
👻 Excluded	804
❓ Unknown	0
🚫 Errors	6
⛔ Unsupported	0

Errors per input

Errors in docs/docs/v2/connectors/mysql.md

• [403] https://dev.mysql.com/doc/connector-net/en/ | Rejected status code: 403 Forbidden (configurable with "accept" option)
• [403] https://dev.mysql.com/doc/connector-net/en/connector-net-connections-string.html | Rejected status code: 403 Forbidden (configurable with "accept" option)

Errors in docs/docs/v3/connectors/mysql.md

• [403] https://dev.mysql.com/doc/connector-net/en/ | Error (cached)
• [403] https://dev.mysql.com/doc/connector-net/en/connector-net-connections-string.html | Error (cached)

Errors in docs/docs/v4/connectors/mysql.md

• [403] https://dev.mysql.com/doc/refman/8.0/en/connecting-using-uri-or-key-value-pairs.html#connection-parameters-base | Rejected status code: 403 Forbidden (configurable with "accept" option)
• [403] https://www.mysql.com/ | Rejected status code: 403 Forbidden (configurable with "accept" option)

Full Github Actions output

[TimHess]

@bart-vmware this one is tricky, I looked for a couple minutes, but it seems like they removed this approach for CosmosDb from the docs?

[bart-vmware]

Yeah, the connection string parameters are mysteriously undocumented by Microsoft. The closest I could find is https://learn.microsoft.com/azure/service-connector/how-to-integrate-cosmos-sql#other-client-types-2. I think the replacement link should be: https://learn.microsoft.com/azure/static-web-apps/database-azure-cosmos-db#get-database-connection-string-for-local-development.

meaningless