docs: cherry-pick documentation reorganization to release/0.6.0

[dagil-nvidia]

docs: cherry-pick documentation reorganization to release/0.6.0

Overview

This PR cherry-picks the documentation reorganization from PR #3658 (commit 598cbbb73) into the release/0.6.0 branch.

The documentation reorganization was merged to main on Oct 17, 2025, and this PR brings those changes to the 0.6.0 release branch.

Replaces: PR #3752 (closed due to base branch conflicts)

Changes

This is a cherry-pick of commit 598cbbb73 which includes:

Directory Reorganization

• docs/API/ → docs/api/ (lowercase for consistency)
• Created docs/development/ for developer-focused guides
• Created docs/observability/ for monitoring, logging, and health checks
• Created docs/performance/ for performance tuning guides
• Created docs/reference/ for CLI reference, glossary, and support matrix
• Moved docs/architecture/kvbm_* → docs/kvbm/ (top-level KV Block Manager docs)
• Moved docs/architecture/planner_* → docs/planner/ (top-level Planner docs)

File Movements (53 files total)

• docs/guides/backend.md → docs/development/backend-guide.md
• docs/runtime/README.md → docs/development/runtime-guide.md
• docs/guides/health_check.md → docs/observability/health-checks.md
• docs/guides/logging.md → docs/observability/logging.md
• docs/guides/metrics.md → docs/observability/metrics.md
• docs/guides/disagg_perf_tuning.md → docs/performance/tuning.md
• docs/guides/dynamo_run.md → docs/reference/cli.md
• docs/dynamo_glossary.md → docs/reference/glossary.md
• docs/support_matrix.md → docs/reference/support-matrix.md
• docs/guides/tool_calling.md → docs/guides/tool-calling.md (renamed)
• docs/architecture/run_kvbm_in_*.md → docs/kvbm/*-setup.md
• docs/components/router/ → docs/router/

Updated References

• Updated all internal documentation links in docs/index.rst
• Updated all references in docs/hidden_toctree.rst
• Updated links in backend READMEs
• Updated relative paths throughout documentation

New Documentation Structure

docs/
├── api/              # API documentation (lowercase)
├── architecture/     # Core architecture docs
├── backends/         # Backend-specific guides
├── benchmarks/       # Benchmarking guides
├── development/      # Developer guides (NEW)
├── guides/           # User guides (REDUCED)
├── kubernetes/       # K8s deployment docs
├── kvbm/            # KV Block Manager docs (NEW)
├── observability/   # Monitoring, logging, health (NEW)
├── performance/     # Performance tuning (NEW)
├── planner/         # Planner-related docs (NEW)
├── reference/       # Reference materials (NEW)
└── router/          # Router documentation

Conflict Resolution

Two simple conflicts were resolved during cherry-pick:

• components/backends/sglang/prometheus.md - Updated metrics.md path from docs/guides/ to docs/observability/
• components/backends/vllm/prometheus.md - Updated metrics.md path from docs/guides/ to docs/observability/

Both conflicts were resolved by accepting the new documentation structure paths.

Testing

• ✅ Cherry-picked cleanly from latest release/0.6.0 branch
• ✅ All moved files exist in new locations
• ✅ All toctree references are valid
• ✅ Internal links updated
• ✅ Conflicts resolved
• ✅ No functional code changes

Related PRs

• Original PR: #3658 (merged to main on Oct 17, 2025)
• Supersedes: PR docs: reorg to release 0.6.0 v2 #3752 (closed - had base branch conflicts)
• Co-authored-by: Anish [email protected]
• Co-authored-by: Claude [email protected]

Checklist

• Cherry-picked from main (commit 598cbbb73)
• Based on latest release/0.6.0
• Conflicts resolved
• Documentation links verified
• No functional code changes
• DCO sign-off maintained from original commit
• Ready to merge without conflicts

[copy-pr-bot]

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.

[github-actions]

📚 Documentation Preview

📖 View Latest Documentation: https://crispy-winner-3jnj38w.pages.github.io/release/0.6.0/pr-3756/

Latest Deployment:

• Commit: 49dbdfc6 (latest push)
• Target: release/0.6.0
• Updated: 10/20/25, 9:57 PM UTC

Note: This link always shows the documentation for the latest commit. It may take a few minutes for GitHub Pages to update after each push.

---

Auto-updated by the documentation deployment workflow