Skip to content

docs: address Harry/VDR feedback + fixing broken links across repository#3802

Merged
saturley-hall merged 37 commits into
mainfrom
harryskim
Oct 22, 2025
Merged

docs: address Harry/VDR feedback + fixing broken links across repository#3802
saturley-hall merged 37 commits into
mainfrom
harryskim

Conversation

@athreesh
Copy link
Copy Markdown
Contributor

@athreesh athreesh commented Oct 22, 2025

Summary

Fixed ~96 broken internal markdown links across 22 files throughout the repository.

Changes Made

Architecture Directory Migration (44 links)

Fixed all references to non-existent docs/architecture/ directory:

  • disagg_serving.mddocs/design_docs/disagg_serving.md
  • kv_cache_routing.mddocs/router/kv_cache_routing.md
  • request_migration.mddocs/fault_tolerance/request_migration.md
  • request_cancellation.mddocs/fault_tolerance/request_cancellation.md
  • architecture.mddocs/design_docs/architecture.md
  • distributed_runtime.mddocs/design_docs/distributed_runtime.md

Kubernetes Documentation (18 links)

Fixed incorrect relative paths in docs/kubernetes/README.md:

  • Updated paths to observability subdirectory
  • Fixed deployment guide references
  • Corrected helm chart and example links

Backend Documentation (12 links)

Updated links in:

  • docs/backends/vllm/README.md
  • docs/backends/trtllm/README.md
  • docs/backends/sglang/README.md
  • components/backends/*/deploy/README.md

Additional Fixes

  • Observability documentation (6 links)
  • Development guides (4 links)
  • CLI reference (3 links)
  • Examples and benchmarks (9 links)

Files Modified

  • Root README.md
  • 3 backend deployment READMEs
  • 3 backend documentation READMEs
  • Kubernetes deployment documentation
  • Observability documentation (metrics, logging, health-checks)
  • Development guides
  • CLI reference
  • Examples (multinode, fault tolerance)
  • And more...

Testing

All links have been verified to point to existing files with correct relative paths.

🤖 Generated with Claude Code

Summary by CodeRabbit

  • Documentation
    • Restructured documentation with improved organization and navigation paths for easier discoverability
    • Added new comprehensive guides for Kubernetes deployment, observability, and multinode setups
    • Introduced user guides for multimodal inference, performance optimization, and SLA planning
    • Reorganized architectural documentation into dedicated design and fault tolerance sections

@athreesh athreesh requested a review from a team as a code owner October 22, 2025 01:17
@athreesh athreesh requested a review from a team October 22, 2025 01:17
@athreesh athreesh requested review from a team as code owners October 22, 2025 01:17
@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot Bot commented Oct 22, 2025

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@athreesh athreesh changed the title Fix broken documentation links across repository docs: address Harry/VDR feedback + fixing broken links across repository Oct 22, 2025
@github-actions github-actions Bot added the docs label Oct 22, 2025
@coderabbitai
Copy link
Copy Markdown
Contributor

coderabbitai Bot commented Oct 22, 2025

Walkthrough

Documentation paths were systematically reorganized across the repository, moving files from an architecture/ folder to more specific categories (design_docs/, router/, fault_tolerance/, kubernetes/observability/). The Sphinx documentation structure was restructured with new section files and a redesigned index. New documentation files were added for observability, deployment guides, and features, while one file was removed.

Changes

Cohort / File(s) Change Summary
Backend Documentation Link Updates
benchmarks/router/README.md, components/backends/trtllm/deploy/README.md, components/backends/vllm/deploy/README.md, components/src/dynamo/router/README.md, docs/backends/sglang/README.md, docs/backends/trtllm/README.md, docs/backends/vllm/README.md
Updated documentation links from architecture/ paths to design_docs/, router/, and fault_tolerance/ paths. KV Cache Routing redirected to docs/router/kv_cache_routing.md; Disaggregated Serving redirected to design_docs/disagg_serving.md; Request Migration redirected to fault_tolerance/request_migration.md.
Kubernetes Documentation Structure and Path Updates
docs/kubernetes/README.md, docs/kubernetes/deployment/create_deployment.md, docs/kubernetes/deployment/minikube.md, docs/kubernetes/dynamo_operator.md, docs/kubernetes/fluxcd.md, docs/kubernetes/grove.md, docs/kubernetes/installation_guide.md
Enhanced documentation structure with improved terminology, expanded examples and guidance, updated relative paths, and refined deployment scenarios. Added SLA-based autoscaling references and improved multi-tenant cluster guidance. Relative paths updated from absolute to facilitate easier navigation.
Kubernetes Observability Documentation
docs/kubernetes/observability/logging.md, docs/kubernetes/observability/metrics.md
Updated relative paths for documentation links and namespace defaults. Adjusted image references and installation navigation to use new relative path structure.
Observability and Related Documentation
docs/observability/health-checks.md, docs/observability/logging.md, docs/observability/metrics.md, deploy/logging/README.md, deploy/metrics/k8s/README.md
Updated documentation links from architecture/ to design_docs/ paths for Distributed Runtime and Architecture overviews. Adjusted Kubernetes observability path references.
Deploy Documentation Link Updates
deploy/inference-gateway/README.md, deploy/cloud/pre-deployment/README.md
Updated KV Cache Routing documentation link from architecture/ to router/ path; adjusted internal navigation anchors and NIXL references.
Examples Documentation Link Updates
examples/README.md, examples/basics/multinode/README.md, examples/fault_tolerance/cancellation/README.md
Updated documentation links to reflect new paths: Disaggregated Serving and KV Cache Routing redirected; Request Cancellation path moved to fault_tolerance/. Added Framework Support section.
Root-Level Documentation Link Updates
README.md, components/README.md, docs/design_docs/architecture.md, docs/development/backend-guide.md, docs/fault_tolerance/request_cancellation.md, docs/reference/cli.md, docs/router/kv_cache_routing.md, docs/backends/sglang/multimodal_epd.md
Updated documentation links from architecture/ to design_docs/, router/, and fault_tolerance/ paths across multiple files. Adjusted component path references and internal link targets.
Sphinx Documentation Structure
docs/_sections/backends.rst, docs/_sections/k8s_deployment.rst, docs/_sections/k8s_multinode.rst, docs/_sections/k8s_observability.rst, docs/_sections/observability.rst
Removed license headers and introductory content from backends section; added new Deployment Guide, Multinode, Observability, and K8s Observability sections with hidden toctree structures linking to relevant documentation.
Documentation Index Restructuring
docs/index.rst
Replaced previous Architectural section and Kubernetes toctree entries with new structure: added Deployment Guide, Observability (K8s), and Multinode sections, and introduced new User Guides toctree containing Tool Calling, Multimodality Support, performance, benchmarking, and observability entries. Replaced Developer Guide caption with Design Docs section.
New Performance and Planner Documentation
docs/performance/aiconfigurator.md, docs/planner/sla_planner.md, docs/planner/sla_planner_quickstart.md
Added new AIConfigurator documentation describing tool for LLM performance optimization with Quick Start and configuration sections. Updated SLA Planner documentation with corrected file path reference and updated profiling guide links.
New Multimodal and Benchmark Documentation
docs/multimodal/multimodal_intro.md, docs/benchmarks/benchmarking.md
Added new multimodal inference documentation describing EPD and PD disaggregation approaches, media handling, and inference framework support matrix. Added "Testing with Mocker Backend" section to benchmarking documentation.
Deleted Documentation
docs/kubernetes/gke_setup.md
Removed comprehensive GKE Workload Identity and Artifact Registry access configuration guide.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

The changes are predominantly homogeneous documentation link migrations and path restructuring distributed across 40+ files. While individually simple, the structural reorganization of the Sphinx documentation index, addition of new documentation files, and the sheer number of affected locations require careful verification to ensure all relative and absolute paths reference correct documentation destinations and no cross-references are broken.

Poem

🐰 Hops through docs with glee,
Paths reorganized, now they're free!
From architecture to homes so new,
Router, fault_tolerance, design_docs too!
Navigation flows like carrots bright,
Documentation's structure: pure delight!

Pre-merge checks

✅ Passed checks (3 passed)
Check name Status Explanation
Description Check ✅ Passed The PR description provides substantial detail about the changes, including a comprehensive summary of the ~96 broken links fixed, well-organized categorization by area (Architecture Directory Migration, Kubernetes Documentation, Backend Documentation, Additional Fixes), specific file modifications, and verification notes. However, the description does not include a dedicated "Where should the reviewer start?" section highlighting specific files for close review, and it lacks a "Related Issues" section with action keywords as specified in the template. Despite these structural deviations, the description is mostly complete with sufficient information to understand the scope and intent of the changes.
Docstring Coverage ✅ Passed No functions found in the changes. Docstring coverage check skipped.
Title Check ✅ Passed The PR title "docs: address Harry/VDR feedback + fixing broken links across repository" accurately reflects the primary objective of the changeset. The raw_summary shows that the core work involves fixing approximately 96 broken internal Markdown links across 22 files, with major corrective work addressing an Architecture directory migration (44 links), Kubernetes documentation paths (18 links), backend documentation updates (12 links), and additional fixes across observability, development guides, CLI reference, and examples. The title is specific enough to convey that documentation links are being corrected, concise at 72 characters, and clearly indicates this is not a localized change but rather a repository-wide fix.

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands and usage tips.

Copy link
Copy Markdown
Contributor

@coderabbitai coderabbitai Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actionable comments posted: 3

🧹 Nitpick comments (5)
docs/multimodal/multimodal_intro.md (1)

26-43: Consider adding cross-references to related architecture documentation.

The EPD and PD descriptions reference concepts like "disaggregation" and "routing" that are likely detailed in other documentation files (based on the PR summary mentioning docs/design_docs/disagg_serving.md and docs/router/kv_cache_routing.md).

Consider enriching this file with brief cross-reference links to related docs:

  • Link to disaggregation architecture details
  • Link to routing/scheduling documentation if applicable

This would improve discoverability and help readers understand the broader system context.

docs/planner/sla_planner.md (1)

105-108: Add language identifiers to code blocks.

Code blocks at lines 105–108 and 112–121 lack language specifiers. While the static analysis hint incorrectly flags line 24 (which is not a code block), these blocks should declare their language for proper syntax highlighting and linter compliance.

Consider adding language identifiers:

-```
+```
 predicted_load = next_requests * next_isl / interval * min(1, prefill_correction)
-```
+```

-```
+```python
 # 1. apply d_correction_factor to the ITL SLA

(Note: if these blocks represent pseudocode or shell, adjust the language identifier accordingly.)

Also applies to: 112-121

components/src/dynamo/router/README.md (1)

6-6: LGTM.

All KV routing documentation links have been updated to reflect the reorganization. Note that the file uses a mix of absolute paths (e.g., /docs/router/, /benchmarks/) and relative paths (e.g., ../vllm/). While both approaches work, standardizing on one approach across all references could improve maintainability.

Also applies to: 32-32, 45-45, 97-97, 101-102

docs/performance/aiconfigurator.md (1)

153-155: Minor: Consider using relative paths for consistency.

The documentation links use absolute paths (/docs/...), which work in GitHub but may not work in local/offline documentation builds. For consistency with other files in this PR, consider using relative paths:

  • /docs/kubernetes/installation_guide.md../kubernetes/installation_guide.md
  • /docs/planner/sla_planner_quickstart.md../planner/sla_planner_quickstart.md
  • /docs/benchmarks/benchmarking.md../benchmarks/benchmarking.md

This is optional for a new file but would improve consistency across documentation.

docs/kubernetes/grove.md (1)

96-100: Minor style note: repeated sentence beginnings.

Lines 96 and 98 both begin with "For", which the style checker flagged. This is a minor style preference rather than a functional issue, but consider rephrasing line 98 to vary the opening if you're doing minor edits. For example: "To see the latest updates on Grove, refer to the [official project on GitHub]..." or similar.

📜 Review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between e6fc0e2 and 10176ac.

📒 Files selected for processing (46)
  • README.md (2 hunks)
  • benchmarks/router/README.md (1 hunks)
  • components/README.md (1 hunks)
  • components/backends/trtllm/deploy/README.md (2 hunks)
  • components/backends/vllm/deploy/README.md (2 hunks)
  • components/src/dynamo/router/README.md (3 hunks)
  • deploy/cloud/pre-deployment/README.md (1 hunks)
  • deploy/inference-gateway/README.md (2 hunks)
  • deploy/logging/README.md (1 hunks)
  • deploy/metrics/k8s/README.md (1 hunks)
  • docs/_sections/backends.rst (1 hunks)
  • docs/_sections/k8s_deployment.rst (1 hunks)
  • docs/_sections/k8s_multinode.rst (1 hunks)
  • docs/_sections/k8s_observability.rst (1 hunks)
  • docs/_sections/observability.rst (1 hunks)
  • docs/backends/sglang/README.md (2 hunks)
  • docs/backends/sglang/multimodal_epd.md (2 hunks)
  • docs/backends/trtllm/README.md (2 hunks)
  • docs/backends/vllm/README.md (3 hunks)
  • docs/benchmarks/benchmarking.md (1 hunks)
  • docs/design_docs/architecture.md (1 hunks)
  • docs/development/backend-guide.md (4 hunks)
  • docs/fault_tolerance/request_cancellation.md (1 hunks)
  • docs/index.rst (2 hunks)
  • docs/kubernetes/README.md (6 hunks)
  • docs/kubernetes/deployment/create_deployment.md (2 hunks)
  • docs/kubernetes/deployment/minikube.md (1 hunks)
  • docs/kubernetes/dynamo_operator.md (1 hunks)
  • docs/kubernetes/fluxcd.md (2 hunks)
  • docs/kubernetes/gke_setup.md (0 hunks)
  • docs/kubernetes/grove.md (1 hunks)
  • docs/kubernetes/installation_guide.md (6 hunks)
  • docs/kubernetes/observability/logging.md (1 hunks)
  • docs/kubernetes/observability/metrics.md (6 hunks)
  • docs/multimodal/multimodal_intro.md (1 hunks)
  • docs/observability/health-checks.md (1 hunks)
  • docs/observability/logging.md (1 hunks)
  • docs/observability/metrics.md (2 hunks)
  • docs/performance/aiconfigurator.md (1 hunks)
  • docs/planner/sla_planner.md (1 hunks)
  • docs/planner/sla_planner_quickstart.md (2 hunks)
  • docs/reference/cli.md (1 hunks)
  • docs/router/kv_cache_routing.md (1 hunks)
  • examples/README.md (1 hunks)
  • examples/basics/multinode/README.md (3 hunks)
  • examples/fault_tolerance/cancellation/README.md (1 hunks)
💤 Files with no reviewable changes (1)
  • docs/kubernetes/gke_setup.md
🧰 Additional context used
🪛 LanguageTool
docs/kubernetes/grove.md

[style] ~97-~97: Three successive sentences begin with the same word. Consider rewording the sentence or use a thesaurus to find a synonym.
Context: ...-node disaggregated serving scenarios. For the latest updates on Grove, refer to t...

(ENGLISH_WORD_REPEAT_BEGINNING_RULE)

docs/kubernetes/installation_guide.md

[style] ~127-~127: Consider an alternative verb to strengthen your wording.
Context: ...paceRestriction.enabled=true`). If you see this validation error, you need namespa...

(IF_YOU_HAVE_THIS_PROBLEM)


[style] ~167-~167: Try using a synonym here to strengthen your writing.
Context: ...tures from the main branch. Note: This gives you access to the latest unreleased feature...

(GIVE_PROVIDE)

🪛 markdownlint-cli2 (0.18.1)
docs/planner/sla_planner.md

24-24: Fenced code blocks should have a language specified

(MD040, fenced-code-language)

🔇 Additional comments (58)
docs/multimodal/multimodal_intro.md (3)

19-21: Use relative path for local examples link.

Given the PR focus on fixing documentation links with correct relative paths, the absolute GitHub URL on line 21 should be a relative path to maintain consistency with the broader documentation structure.

Please verify:

  1. Does examples/multimodal/ exist in the repository?
  2. Should this link be [multimodal examples](../../examples/multimodal) instead of the absolute GitHub URL?

If the intent is to link to deployed web documentation, keep the absolute URL. Otherwise, convert to a relative path for internal consistency.


45-53: Verify table alignment and rendering.

The support matrix table uses emoji indicators (✅, 🚧, ❌) for visual clarity. Ensure the table renders correctly in your Sphinx/Markdown renderer of choice, particularly the vertical bar separators and alignment.

Render this documentation locally to confirm the table displays correctly with proper column alignment and emoji rendering.


1-17: Licensing header is properly formatted.

The SPDX-compliant Apache-2.0 license header follows the correct format and aligns with repository conventions.

docs/_sections/backends.rst (1)

7-9: RST toctree formatting is consistent and correct.

The backend entries now follow uniform syntax with proper label-path formatting and consistent indentation. All relative paths reference ../backends/{name}/README, which correctly resolves from the docs/_sections/ location to docs/backends/.

The removal of inconsistent formatting (noted in the summary regarding the TensorRT-LLM dash adjustment) brings this section into alignment with reStructuredText best practices for toctree directives.

docs/planner/sla_planner.md (1)

24-24: Documentation link path updated correctly as part of repository restructuring.

The link path has been updated from /components/planner/src/dynamo/planner/defaults.py to /components/src/dynamo/planner/defaults.py, aligning with the broader documentation reorganization described in the PR objectives. Per the PR summary, all links have been verified to point to existing files with correct relative paths.

docs/reference/cli.md (1)

155-155: All documentation links verified successfully.

The three link updates in docs/reference/cli.md (lines 155, 161, 167) all resolve correctly to existing files:

  • ../fault_tolerance/request_migration.md
  • ../fault_tolerance/request_cancellation.md

The migration from the architecture/ directory to fault_tolerance/ paths is complete and valid.

deploy/cloud/pre-deployment/README.md (1)

24-24: LGTM.

Linking to the main branch version provides users with clear access to the latest documentation.

docs/fault_tolerance/request_cancellation.md (1)

48-48: LGTM.

Link correctly reflects the relocation of the cancellation demo to the fault_tolerance examples directory.

deploy/metrics/k8s/README.md (1)

3-3: LGTM.

Link correctly reflects the reorganization of Kubernetes observability content under a dedicated observability subdirectory.

examples/fault_tolerance/cancellation/README.md (1)

91-91: LGTM.

Link correctly reflects the movement of the cancellation documentation to the fault_tolerance documentation directory.

docs/router/kv_cache_routing.md (1)

213-213: LGTM.

Link correctly uses absolute path to reference the reorganized distributed runtime documentation in the design_docs directory.

components/README.md (1)

34-34: Link path update verified—no issues found.

The backends/ directory exists at components/backends/ and contains backend-specific documentation organized in subdirectories (sglang, trtllm, vllm). The relative link reference correctly resolves to the target directory structure.

docs/_sections/k8s_multinode.rst (1)

1-8: No issues found—Sphinx configuration correctly resolves the toctree paths.

Both referenced files exist (docs/kubernetes/deployment/multinode-deployment.md and docs/kubernetes/grove.md), and your Sphinx configuration is properly set up with myst_parser and source_suffix = [".rst", ".md"], which automatically handles markdown file resolution without requiring explicit extensions in toctree entries. The relative paths will resolve correctly.

benchmarks/router/README.md (1)

119-119: Link path update is correct.

The relative path ../../docs/router/kv_cache_routing.md correctly resolves from benchmarks/router/README.md to the new KV Cache Routing documentation location. Link format and path arithmetic are correct.

deploy/inference-gateway/README.md (2)

163-163: Link path update is correct.

The relative path ../../docs/router/kv_cache_routing.md correctly resolves from deploy/inference-gateway/README.md to the new KV Cache Routing documentation location.


19-19: Anchor link #6-usage is correctly implemented.

Line 257 contains the heading ### 6. Usage ###, which generates the anchor #6-usage in markdown. The table of contents reference on line 19 is valid and will resolve correctly. No action needed.

components/backends/vllm/deploy/README.md (2)

227-227: Link path update for request migration documentation is correct.

The relative path ../../../../docs/fault_tolerance/request_migration.md correctly resolves from the nested file location components/backends/vllm/deploy/README.md to the new request migration documentation.


242-242: Link paths for architecture documentation are correct.

The relative paths for Disaggregated Serving (../../../../docs/design_docs/disagg_serving.md) and KV-Aware Routing (../../../../docs/router/kv_cache_routing.md) correctly resolve from the nested file location with proper relative path depth.

docs/development/backend-guide.md (4)

77-77: Link path update for request migration parameter documentation is correct.

The relative path ../fault_tolerance/request_migration.md correctly resolves from docs/development/backend-guide.md to the new request migration documentation location.


119-119: Link path update for request migration generics documentation is correct.

The relative path ../fault_tolerance/request_migration.md correctly resolves for the GeneratorExit exception documentation reference.


142-142: Link path update for request migration architecture documentation is correct.

The relative path ../fault_tolerance/request_migration.md correctly resolves to the new request migration architecture documentation location.


164-164: Request cancellation documentation link is valid.

The file docs/fault_tolerance/request_cancellation.md exists at the expected location, confirming the reference in docs/development/backend-guide.md (line 164) is correct and the link will not break.

docs/design_docs/architecture.md (1)

55-56: No issues found. Both relative paths are correct and consistent.

The verification confirms both referenced files exist at their expected locations. Line 55's disagg_serving.md correctly resolves to docs/design_docs/disagg_serving.md (same directory), and line 56's ../router/kv_cache_routing.md correctly resolves to docs/router/kv_cache_routing.md. The paths follow consistent relative linking conventions based on file locations.

Likely an incorrect or invalid review comment.

docs/planner/sla_planner_quickstart.md (2)

18-18: The review comment is based on an incorrect premise.

The actual file uses GitHub's standard admonition syntax > [!TIP], not the -> or +> markers mentioned in the review comment. The file consistently uses this syntax throughout (> [!IMPORTANT], > [!TIP], > [!WARNING]), and no files in the repository use +> or -> patterns. The concern about the syntax change is unfounded.

Likely an incorrect or invalid review comment.


41-41: No action required; line 41 is consistent with the file's existing link format pattern.

The verification confirms that docs/kubernetes/observability/metrics.md exists and the absolute path is valid. More importantly, the original review's premise is incorrect: line 41 is not inconsistent with other files in the PR. Within sla_planner_quickstart.md, absolute paths are the standard format used across most links (lines 6, 18, 40, 41, 145, 248, 252–254), with only line 42 using a relative path. Sphinx documentation systems handle absolute internal links correctly, and the path format poses no build issues.

docs/_sections/k8s_deployment.rst (1)

4-10: All referenced Kubernetes documentation files exist and are correctly linked.

Verification confirmed that all four toctree references in the file point to existing documentation files:

  • docs/kubernetes/README.md
  • docs/kubernetes/installation_guide.md
  • docs/kubernetes/dynamo_operator.md
  • docs/kubernetes/deployment/minikube.md

The relative paths resolve correctly from docs/_sections/ to the Kubernetes documentation directory, and Sphinx will properly auto-detect the .md extensions. No issues found.

README.md (1)

59-61: Confirm whether absolute paths work in your documentation context.

The referenced files and anchor all exist:

  • docs/design_docs/disagg_serving.md
  • docs/router/kv_cache_routing.md
  • #conditional-disaggregation anchor exists at line 52

However, absolute paths like /docs/... may not resolve correctly when the README is viewed directly on GitHub.com. If this README is primarily viewed on GitHub (not through a docs build system), consider using relative paths (./docs/... or docs/...) instead for better portability.

Verify these links work as intended in your documentation viewing context before merging.

components/backends/trtllm/deploy/README.md (2)

261-261: Approved: Link update aligns with documentation restructuring.

The request migration link correctly updates from architecture/ to fault_tolerance/ with proper relative path (../../../../docs/fault_tolerance/request_migration.md).


284-284: Approved: Architecture documentation links properly updated.

Links to disaggregated serving and KV-aware routing correctly migrate from docs/architecture/ to docs/design_docs/ and docs/router/ with correct relative paths.

docs/backends/trtllm/README.md (2)

55-57: Approved: Feature matrix links updated correctly.

Architecture documentation links in the feature support matrix correctly migrate to new structure:

  • Disaggregated Serving: design_docs/disagg_serving.md
  • Conditional Disaggregation: design_docs/disagg_serving.md#conditional-disaggregation
  • KV-Aware Routing: router/kv_cache_routing.md

Relative paths are correctly formatted with ../../../docs/ prefix.


223-229: Approved: Request migration links updated to fault_tolerance.

The request migration references correctly migrate from architecture/request_migration.md to fault_tolerance/request_migration.md with proper relative paths (../../../docs/fault_tolerance/request_migration.md).

docs/kubernetes/fluxcd.md (1)

3-3: Approved: Cross-document links converted to relative paths.

Links properly converted from absolute paths to relative paths:

  • Aggregated vLLM example: /docs/backends/vllm/README.md../backends/vllm/README.md
  • Installation guide references: /docs/kubernetes/installation_guide.md./installation_guide.md

Relative paths are more maintainable and work correctly in different documentation contexts.

Also applies to: 7-7, 21-21

docs/backends/sglang/README.md (1)

37-39: Approved: Documentation links restructured consistently.

All links updated in alignment with documentation restructuring:

  • Feature matrix entries migrate from architecture/ to design_docs/, router/ paths
  • Migration documentation link: fault_tolerance/request_migration.md

Relative paths use correct prefix ../../ for docs/backends/sglang/ → docs/* navigation.

Also applies to: 58-58

examples/README.md (1)

33-40: Approved: Content organization improvements.

The Framework Support section addition and deployment examples reorganization improve documentation discoverability. The reworded section header ("Runtime Examples" → "Low-level runtime examples") clarifies the purpose of those examples.

Also applies to: 48-48, 52-54

docs/backends/vllm/README.md (2)

38-40: Approved: Backend documentation links restructured correctly.

Feature matrix and Kubernetes deployment guide links properly updated:

  • Architecture documentation migrated to design_docs/ and router/ with correct relative paths
  • Kubernetes deployment link correctly points to updated location

Also applies to: 156-156


181-181: Approved: Request migration and KV routing references updated.

Both KV cache routing and request migration documentation references correctly migrate to new paths:

  • KV Cache Routing: ../../../docs/router/kv_cache_routing.md
  • Request Migration: ../../../docs/fault_tolerance/request_migration.md

Relative path prefixes are consistent with file location (docs/backends/vllm/).

Also applies to: 185-185, 229-229

docs/kubernetes/installation_guide.md (2)

244-246: Approved: Backend deployment guide links correctly updated.

Links to backend deployment READMEs use proper relative paths from docs/kubernetes/ to backend locations:

  • vLLM: ../../components/backends/vllm/deploy/README.md
  • SGLang: ../../components/backends/sglang/deploy/README.md
  • TensorRT-LLM: ../../components/backends/trtllm/deploy/README.md

Path traversal is correct (up 2 levels from docs/kubernetes/ to repository root).


313-315: Approved: Internal documentation links properly formatted.

Links to Helm chart configuration, deployment guides, and operator documentation use correct relative paths:

  • ../../deploy/cloud/helm/platform/README.md
  • ./deployment/create_deployment.md
  • ./dynamo_operator.md

Navigation structure is clear and maintainable.

docs/_sections/observability.rst (1)

1-9: Structure and syntax look correct.

The Sphinx toctree links are properly formatted with correct relative paths. This integrates cleanly with the observability documentation reorganization.

Can you verify that the three target files exist at these locations: docs/observability/metrics.md, docs/observability/logging.md, and docs/observability/health-checks.md?

docs/_sections/k8s_observability.rst (1)

1-8: Structure and syntax look correct.

Sphinx toctree links are properly formatted for the Kubernetes observability section with accurate relative paths.

Can you verify that the two target files exist: docs/kubernetes/observability/metrics.md and docs/kubernetes/observability/logging.md?

docs/kubernetes/observability/logging.md (1)

20-20: Reference path is correct.

The relative path to the Kubernetes README is properly formatted.

deploy/logging/README.md (1)

3-3: Relative path is correct and points to the right location.

From deploy/logging/README.md, the path correctly traverses to docs/kubernetes/observability/logging.md.

docs/kubernetes/deployment/minikube.md (1)

61-61: Path correction aligns with documentation structure reorganization.

Updating from ./installation_guide.md to ../installation_guide.md correctly reflects that the installation guide is now at the parent directory level.

docs/kubernetes/grove.md (1)

96-100: Relative paths are correct.

Both link updates properly reference the documentation structure: multinode deployment guide under ./deployment/ and installation guide at the kubernetes directory level.

docs/observability/health-checks.md (1)

198-199: Architecture documentation links correctly updated to new design_docs paths.

Both links properly reflect the migration from ../architecture/ to ../design_docs/, which aligns with the PR's systematic reorganization of architecture documentation.

Can you verify that both target files exist: docs/design_docs/distributed_runtime.md and docs/design_docs/architecture.md?

docs/observability/logging.md (1)

188-189: Architecture documentation links correctly updated to new design_docs paths.

Both links follow the same migration pattern: moving from ../architecture/ to ../design_docs/. The paths are structurally sound.

Can you verify that both target files exist: docs/design_docs/distributed_runtime.md and docs/design_docs/architecture.md?

examples/basics/multinode/README.md (1)

7-8: Link updates look consistent and well-coordinated.

Both links use identical relative path structure (../../../docs/) and target files in the new documentation hierarchy (design_docs and router directories). The updates are consistent across all three references in the file.

docs/kubernetes/observability/metrics.md (1)

31-31: Relative paths are correctly updated for new file locations.

The links properly reference installation_guide.md in the parent directory, and cross-references to backend and observability documentation use appropriate relative path traversal (../../) from the observability subdirectory.

Also applies to: 67-68, 90-90

docs/observability/metrics.md (1)

34-34: Documentation path reorganization is properly reflected.

Cross-references to kubernetes/observability and design_docs directories are correctly updated. Relative paths (../) properly traverse from docs/observability/ to peer directories in the docs hierarchy.

Also applies to: 97-98

docs/kubernetes/deployment/create_deployment.md (2)

31-31: Verify other architecture pattern example links for consistency.

Lines 46, 65, 73, 75, and 77 contain similar example links to deployment templates. Ensure these links have complete URLs on single lines without breaks.

Also applies to: 46-46, 65-65, 73-73, 75-75, 77-77


148-148: CLI reference link path updated correctly.

The reference to ../../reference/cli.md is appropriately updated for the file's location in docs/kubernetes/deployment/.

docs/kubernetes/dynamo_operator.md (2)

26-26: Relative path reference to API documentation correctly updated.

The link to ./api_reference.md properly references a sibling file in the same directory.


30-75: Installation instructions significantly enhanced with clear alternatives.

The restructuring provides both Quick Install (via Helm) and Building from Source options with environment variables, making it easier for users to choose their installation path. The addition at line 76 linking to the Installation Guide provides an appropriate reference for detailed customization options.

docs/index.rst (1)

39-86: Documentation navigation structure redesigned with clear organizational logic.

The restructuring effectively separates concerns:

  • Kubernetes deployment/observability topics in dedicated section (lines 52-54)
  • User guides consolidated (tool calling, multimodality, performance, benchmarking)
  • Design documentation highlighted in separate section (lines 83-86)

This provides clearer navigation compared to mixing architecture and guides. The renaming from "Developer Guide" to "Design Docs" better reflects the content focus.

Verify that the referenced section files exist in the new structure:

  • _sections/k8s_deployment
  • _sections/k8s_observability
  • _sections/k8s_multinode
  • _sections/observability

And design docs exist at:

  • design_docs/architecture.md
  • design_docs/dynamo_flow.md
  • design_docs/disagg_serving.md
  • design_docs/distributed_runtime.md
docs/kubernetes/README.md (4)

22-33: New "Important Terminology" section clarifies important distinction.

The distinction between Kubernetes Namespace and Dynamo Namespace is well-explained and valuable for users. This addition helps prevent a common source of confusion in multi-tenant or complex deployments.


62-62: Clarify relative path reference to installation guide.

Line 62 references the installation guide with path ../docs/kubernetes/installation_guide.md. Since this file is already at docs/kubernetes/README.md, this path appears unnecessary (traversing up to docs level then back down). This should likely be ./installation_guide.md (sibling file).

Verify the correct relative path:

  • If the reference should point to docs/kubernetes/installation_guide.md, use ./installation_guide.md
  • If it should point elsewhere, clarify the intended target

64-72: Backend deployment table links properly formatted.

The table uses correct relative paths (../../components/backends/) to reference backend README files for SGLang, TensorRT-LLM, and vLLM backends.


218-226: Additional Resources section links require verification.

Several links in the Additional Resources section use ../docs/ paths (e.g., line 218-219). Given that this file is at docs/kubernetes/README.md, these paths may be incorrect. Verify that paths like ../docs/kubernetes/... should instead be ./... (sibling) or ../... (peer directory).

Comment thread docs/backends/sglang/multimodal_epd.md Outdated
Comment thread docs/benchmarks/benchmarking.md Outdated
Comment thread docs/kubernetes/deployment/create_deployment.md Outdated
harryskim and others added 16 commits October 21, 2025 18:24
Signed-off-by: Harry Kim <harry_kim@live.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Add design doc at the end and remove architecture from getstarted.

Signed-off-by: Harry Kim <harry_kim@live.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Fixed ~96 broken internal markdown links across 22 files:

- Fixed missing docs/architecture/ directory references (44 links)
  - Updated to correct locations: docs/design_docs/, docs/router/, docs/fault_tolerance/
- Fixed incorrect relative paths in docs/kubernetes/README.md (18 links)
- Fixed backend deployment documentation links (12 links)
- Fixed observability documentation links (6 links)
- Fixed miscellaneous path issues (16 links)

Key changes:
- docs/architecture/disagg_serving.md → docs/design_docs/disagg_serving.md
- docs/architecture/kv_cache_routing.md → docs/router/kv_cache_routing.md
- docs/architecture/request_migration.md → docs/fault_tolerance/request_migration.md
- docs/kubernetes/metrics.md → docs/kubernetes/observability/metrics.md
- Updated all relative paths to match actual file locations

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
added multimodality docs

Signed-off-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Signed-off-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
@athreesh athreesh enabled auto-merge (squash) October 22, 2025 16:53
athreesh and others added 4 commits October 22, 2025 13:32
Updated test path from examples/custom_backend/cancellation to
examples/fault_tolerance/cancellation to reflect the example's
new location.

Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Move examples/fault_tolerance/cancellation back to
examples/custom_backend/cancellation to match main branch
structure. Updated all references in tests and documentation.

Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Updated references from docs/guides/metrics.md to docs/observability/metrics.md
to reflect the documentation reorganization.

Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
@athreesh
Copy link
Copy Markdown
Contributor Author

/ok to test d5d447b

@nv-anants
Copy link
Copy Markdown
Member

/ok to test bd06e4d

Signed-off-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Signed-off-by: Harrison King Saturley-Hall <hsaturleyhal@nvidia.com>
@saturley-hall saturley-hall merged commit c6b5904 into main Oct 22, 2025
12 checks passed
@saturley-hall saturley-hall deleted the harryskim branch October 22, 2025 23:56
saturley-hall added a commit that referenced this pull request Oct 23, 2025
…ory (#3802)

Signed-off-by: Harry Kim <harry_kim@live.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Signed-off-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Signed-off-by: Harrison King Saturley-Hall <hsaturleyhal@nvidia.com>
Co-authored-by: Harry Kim <harry_kim@live.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Co-authored-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Signed-off-by: Harrison King Saturley-Hall <hsaturleyhal@nvidia.com>
saturley-hall added a commit that referenced this pull request Oct 23, 2025
…ory (#3802) (#3841)

Signed-off-by: Harry Kim <harry_kim@live.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Signed-off-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Signed-off-by: Harrison King Saturley-Hall <hsaturleyhal@nvidia.com>
Co-authored-by: Anish <80174047+athreesh@users.noreply.github.com>
Co-authored-by: Harry Kim <harry_kim@live.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
This was referenced Oct 29, 2025
yao531441 pushed a commit to yao531441/dynamo that referenced this pull request May 13, 2026
…ory (ai-dynamo#3802)

Signed-off-by: Harry Kim <harry_kim@live.com>
Signed-off-by: athreesh <anish.maddipoti@utexas.edu>
Signed-off-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Signed-off-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Signed-off-by: Harrison King Saturley-Hall <hsaturleyhal@nvidia.com>
Co-authored-by: Harry Kim <harry_kim@live.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: akshatha-k <33278067+akshatha-k@users.noreply.github.com>
Co-authored-by: Harrison Saturley-Hall <hsaturleyhal@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

9 participants