{ "schema": "fak-study-inventory-map/1", "repository": "llm-d/llm-d", "url": "https://github.com/llm-d/llm-d", "indexed_revision": "3243fcf1191348b55c7811267a98117f8b7a6910", "observed_at": "2026-08-25T19:00:00-07:00", "totals": { "files": 1129, "directories": 593, "bytes": 30601578, "runtime_files": 771, "test_files": 3, "docs_files": 274, "text_lines": 101395 }, "source_classes": [ { "class": "readme_docs", "status": "covered", "evidence": [ ".github/scripts/nightly-e2e-verification/README.md", ".github/workflows/README.md", "README.md", "docs/api-reference/README.md", "docs/api-reference/artifacts.md", "docs/api-reference/endpointpickerconfig.md", "docs/api-reference/epp-grpc-apis.md", "docs/api-reference/epp-http-apis.md", "docs/api-reference/epp-http-headers.md", "docs/api-reference/glossary.md", "docs/api-reference/inferencemodelrewrite.md", "docs/api-reference/inferenceobjective.md" ], "note": "local tree evidence found" }, { "class": "architecture_design", "status": "covered", "evidence": [ "docs/architecture/README.md", "docs/architecture/advanced/autoscaling/README.md", "docs/architecture/advanced/autoscaling/hpa-architecture.svg", "docs/architecture/advanced/autoscaling/hpa-epp.md", "docs/architecture/advanced/autoscaling/hpa-wva.md", "docs/architecture/advanced/autoscaling/scaling-engine-pipeline.svg", "docs/architecture/advanced/autoscaling/slo-aware-keda.md", "docs/architecture/advanced/autoscaling/wva-architecture.svg", "docs/architecture/advanced/batch/README.md", "docs/architecture/advanced/batch/async-processor.md", "docs/architecture/advanced/batch/batch-gateway.md", "docs/architecture/advanced/disaggregation/README.md" ], "note": "local tree evidence found" }, { "class": "runtime_source", "status": "covered", "evidence": [ ".github/ISSUE_TEMPLATE/bug.yml", ".github/ISSUE_TEMPLATE/feature-request.yml", ".github/ISSUE_TEMPLATE/hardware-platform-support.yml", ".github/actionlint.yaml", ".github/dependabot.yml", ".github/scripts/e2e/e2e-validate-flow-control.sh", ".github/scripts/e2e/e2e-validate-predicted-latency.sh", ".github/scripts/e2e/e2e-validate-scale-event.sh", ".github/scripts/e2e/e2e-validate.sh", ".github/scripts/e2e/helm-get-all.sh", ".github/scripts/e2e/wide-ep-transform.sh", ".github/scripts/install-agentgateway-crds.sh" ], "note": "local tree evidence found" }, { "class": "tests_fixtures", "status": "covered", "evidence": [ ".github/scripts/nightly-e2e-verification/test_verify_helpers.py", ".github/scripts/nightly-e2e-verification/tiered-prefix-cache/test_verify.py", "guides/batch-serving/asynchronous-processing/test/kind-e2e.sh" ], "note": "local tree evidence found" }, { "class": "history_changelog_releases", "status": "covered", "evidence": [ ".github/ISSUE_TEMPLATE/new-release.md", ".github/workflows/ci-release.yaml", ".github/workflows/release-matrix.yaml", ".github/workflows/verify-release-images.yaml", "scripts/sync-release-matrix.py" ], "note": "local tree evidence found" }, { "class": "open_closed_issues_prs_discussions", "status": "partial", "evidence": [ ".github/ISSUE_TEMPLATE/bug.yml", ".github/ISSUE_TEMPLATE/feature-request.yml", ".github/ISSUE_TEMPLATE/hardware-platform-support.yml", ".github/ISSUE_TEMPLATE/new-release.md" ], "note": "local templates found; open/closed issue, PR, and discussion history still require GitHub or forge read-back" }, { "class": "roadmap_todos", "status": "checked_absent", "note": "local tree scanned and no matching path was found" }, { "class": "license_provenance", "status": "covered", "evidence": [ "LICENSE" ], "note": "local tree evidence found" }, { "class": "fak_selfquery_witness", "status": "external_required", "note": "process artifact created by the study pass, not by the foreign repository tree" }, { "class": "candidate_matrix", "status": "external_required", "note": "process artifact created by the study pass, not by the foreign repository tree" }, { "class": "completeness_critic", "status": "covered", "note": "generated by this inventory map's local tree walk; non-tree omissions remain named in the note" }, { "class": "issue_tracking", "status": "external_required", "note": "process artifact created by the study pass, not by the foreign repository tree" } ], "subsystems": [ { "path": "guides", "files": 739, "directories": 506, "bytes": 8984258, "runtime_files": 593, "test_files": 1, "docs_files": 94, "text_lines": 55806, "languages": [ { "language": "Config", "files": 571 }, { "language": "Docs", "files": 94 }, { "language": "Shell", "files": 20 }, { "language": "Python", "files": 3 } ], "example_paths": [ "guides/README.md", "guides/agentic-serving/README.md", "guides/agentic-serving/benchmark-templates/agentic-serving-nemotron-3-ultra.yaml", "guides/agentic-serving/benchmark-templates/guide.yaml", "guides/agentic-serving/glm-5-2-h200.md", "guides/agentic-serving/modelserver/gpu/vllm/glm-5-2/kustomization.yaml", "guides/agentic-serving/modelserver/gpu/vllm/nemotron-3-ultra/base/kustomization.yaml", "guides/agentic-serving/modelserver/gpu/vllm/nemotron-3-ultra/base/patch-decode.yaml" ], "source_classes": [ "architecture_design", "readme_docs", "runtime_source", "tests_fixtures" ] }, { "path": ".github", "files": 139, "directories": 7, "bytes": 526447, "runtime_files": 130, "test_files": 2, "docs_files": 6, "text_lines": 13707, "languages": [ { "language": "Config", "files": 113 }, { "language": "Shell", "files": 14 }, { "language": "Docs", "files": 6 }, { "language": "Python", "files": 5 } ], "example_paths": [ ".github/ISSUE_TEMPLATE/bug.yml", ".github/ISSUE_TEMPLATE/feature-request.yml", ".github/ISSUE_TEMPLATE/hardware-platform-support.yml", ".github/ISSUE_TEMPLATE/new-release.md", ".github/actionlint.yaml", ".github/dependabot.yml", ".github/scripts/e2e/e2e-validate-flow-control.sh", ".github/scripts/e2e/e2e-validate-predicted-latency.sh" ], "source_classes": [ "history_changelog_releases", "open_closed_issues_prs_discussions", "readme_docs", "runtime_source", "tests_fixtures" ] }, { "path": "docs", "files": 136, "directories": 38, "bytes": 10493121, "runtime_files": 0, "test_files": 0, "docs_files": 136, "text_lines": 21715, "languages": [ { "language": "Docs", "files": 91 }, { "language": "Config", "files": 9 }, { "language": "Shell", "files": 2 } ], "example_paths": [ "docs/api-reference/README.md", "docs/api-reference/artifacts.md", "docs/api-reference/endpointpickerconfig.md", "docs/api-reference/epp-grpc-apis.md", "docs/api-reference/epp-http-apis.md", "docs/api-reference/epp-http-headers.md", "docs/api-reference/glossary.md", "docs/api-reference/inferencemodelrewrite.md" ], "source_classes": [ "architecture_design", "readme_docs" ] }, { "path": "docker", "files": 37, "directories": 21, "bytes": 10042553, "runtime_files": 16, "test_files": 0, "docs_files": 3, "text_lines": 1299, "languages": [ { "language": "Shell", "files": 13 }, { "language": "Docs", "files": 3 }, { "language": "Config", "files": 2 }, { "language": "Python", "files": 1 } ], "example_paths": [ "docker/build-constraints.txt", "docker/constraints.txt", "docker/packages/common/runtime-otel-package-requirements.txt", "docker/packages/cuda/builder-packages.json", "docker/packages/cuda/runtime-packages.json", "docker/scripts/common/build-uccl.sh", "docker/scripts/common/setup-sccache.sh", "docker/scripts/cpu/install_nixl.py" ], "source_classes": [ "runtime_source" ] }, { "path": ".", "files": 26, "directories": 10, "bytes": 120760, "runtime_files": 6, "test_files": 0, "docs_files": 13, "text_lines": 1894, "languages": [ { "language": "Docs", "files": 13 }, { "language": "Config", "files": 6 } ], "example_paths": [ ".hadolint.yaml", ".markdownlint.json", ".pre-commit-config.yaml", ".prowlabels.yaml", "ADOPTERS.md", "CODE_OF_CONDUCT.md", "CONTRIBUTING.md", "LICENSE" ], "source_classes": [ "license_provenance", "readme_docs", "runtime_source" ] }, { "path": "helpers", "files": 22, "directories": 10, "bytes": 72590, "runtime_files": 18, "test_files": 0, "docs_files": 4, "text_lines": 1791, "languages": [ { "language": "Config", "files": 15 }, { "language": "Docs", "files": 4 }, { "language": "Shell", "files": 3 } ], "example_paths": [ "helpers/benchmark.md", "helpers/client-setup/README.md", "helpers/client-setup/install-deps.sh", "helpers/hf-token.md", "helpers/interactive-pod/manifests/deployment.yaml", "helpers/interactive-pod/manifests/kustomization.yaml", "helpers/interactive-pod/manifests/rbac.yaml", "helpers/interactive-pod/manifests/sa.yaml" ], "source_classes": [ "readme_docs", "runtime_source" ] }, { "path": "proposals", "files": 17, "directories": 1, "bytes": 251217, "runtime_files": 0, "test_files": 0, "docs_files": 16, "text_lines": 2705, "languages": [ { "language": "Docs", "files": 16 } ], "example_paths": [ "proposals/PROPOSAL_TEMPLATE.md", "proposals/autoscaler.md", "proposals/batch-gateway.md", "proposals/coordinator.md", "proposals/distributed-tracing.md", "proposals/inference-resilience-operator.md", "proposals/llm-d-async.md", "proposals/llm-d-inference-payload-processor.md" ] }, { "path": "scripts", "files": 9, "directories": 0, "bytes": 80313, "runtime_files": 8, "test_files": 0, "docs_files": 1, "text_lines": 2275, "languages": [ { "language": "Python", "files": 6 }, { "language": "Shell", "files": 2 }, { "language": "Docs", "files": 1 } ], "example_paths": [ "scripts/ENVVARS.md", "scripts/guide.py", "scripts/lint-dockerfile-envvars.py", "scripts/lint-envvars.py", "scripts/matrix_common.py", "scripts/sync-nightly-matrix.py", "scripts/sync-release-matrix.py", "scripts/test-dockerfile-lint.sh" ], "source_classes": [ "history_changelog_releases", "runtime_source" ] }, { "path": "patches", "files": 2, "directories": 0, "bytes": 2637, "runtime_files": 0, "test_files": 0, "docs_files": 0, "text_lines": 0 }, { "path": "hooks", "files": 1, "directories": 0, "bytes": 147, "runtime_files": 0, "test_files": 0, "docs_files": 0, "text_lines": 0 }, { "path": "release", "files": 1, "directories": 0, "bytes": 27535, "runtime_files": 0, "test_files": 0, "docs_files": 1, "text_lines": 203, "languages": [ { "language": "Docs", "files": 1 } ], "example_paths": [ "release/README.md" ], "source_classes": [ "readme_docs" ] } ], "skipped_dirs": [ ".git", "helpers/interactive-pod/build" ], "completeness_critic": "local tree inventory walked every non-skipped regular file and grouped immediate subsystems; skipped dependency/cache/control directories: .git, helpers/interactive-pod/build; classes checked absent in local tree: roadmap_todos; still requires non-tree study artifacts: fak_selfquery_witness, candidate_matrix, issue_tracking", "non_tree_study": { "observed_at": "2026-08-25T19:00:00-07:00", "source_cutoff": "Pinned tree revision 3243fcf1191348b55c7811267a98117f8b7a6910 (2026-08-26T01:42:53Z); GitHub metadata observed 2026-08-25 America/Los_Angeles.", "github_surface": { "issues": { "total": 533, "open": 124, "closed": 409 }, "pull_requests": { "total": 1841, "open": 100, "closed_unmerged": 472, "merged": 1269 }, "discussions": { "total": 0 }, "releases": 11, "tags": 66, "branches": 11, "milestones": 11, "projects": 1, "latest_release": "v0.9.0 (2026-08-17)", "next_release_signals": [ "milestone/16 v0.10.0", "milestone/14 v1.0.0", "milestone/15 v0.9.0 remains open after release" ], "history": { "commits_since_prior_check": 17, "prior_revision": "48fa8c0a76012038c76f30936fff9727c2e6909d", "pinned_commit": "3243fcf1191348b55c7811267a98117f8b7a6910" }, "method": "GitHub GraphQL exact aggregate counts; paginated REST issues, releases, tags, branches, milestones, and projects; gh pr list read-back; GraphQL discussions; local clone history and refs." }, "roadmap_todos": { "tree_scan": "Case-insensitive TODO/FIXME/roadmap/future-work/planned/follow-up scan across the pinned tree; no dedicated roadmap file. Generated dependency TODOs were excluded from product-roadmap evidence.", "forge_signals": [ "v0.10.0 milestone", "v1.0.0 milestone", "open guides-definition PR #2349" ], "decision": "Treat milestones and active proposals as directional signals, not shipped contracts." }, "license_provenance": { "spdx": "Apache-2.0", "path": "LICENSE", "repository_api": "license.spdx_id=Apache-2.0", "decision": "Borrow ideas or adapted contracts with source attribution; do not copy implementation without retaining applicable notices and license provenance." }, "self_query_axes": [ { "query": "fak capabilities \"Kubernetes distributed inference endpoint routing KV cache locality flow control prefill decode disaggregation autoscaling\"", "result": "Native model routing, turn-tax, stable-context reuse, live control, and savings attribution surfaced; no Kubernetes inference control plane equivalent." }, { "query": "fak capabilities \"llm-d inference scheduler endpoint picker KV events prefix cache compatibility\"", "result": "Native context reuse and guarded fleet surfaces surfaced; the llm-d endpoint-picker and KV-event protocol remain external integration boundaries." }, { "query": "fak capabilities \"vLLM Kubernetes gateway inference extension workload variant autoscaling\"", "result": "Model routing surfaced, but no reason to duplicate Gateway API, Kubernetes autoscaling, or llm-d deployment assembly inside fak." } ], "candidate_matrix": [ { "candidate": "llm-d v0.9 ownership and double-routing contract", "source": [ "docs/architecture/README.md", "docs/api-reference/endpointpickerconfig.md", "guides/" ], "alternative": "Undocumented composition between fak and llm-d", "classification": "adaptation", "disposition": "already_owned", "reason": "FAK issue #8016 shipped docs/integrations/llm-d.md. Keep fak authoritative for policy/model choice and llm-d authoritative for cluster endpoint selection." }, { "candidate": "pinned live-cluster compatibility witness", "source": [ "guides/", "docs/architecture/", "v0.9.0 release" ], "alternative": "Static documentation-only interoperability claim", "classification": "integration_witness", "disposition": "follow_on_open", "reason": "FAK issue #8017 owns the live compatibility proof; this inventory refresh does not substitute for a cluster receipt." }, { "candidate": "KV-event-informed prefix locality and speculative cache-warm metadata", "source": [ "docs/architecture/", "docs/architecture/advanced/disaggregation/" ], "alternative": "FAK-native prefix reuse without llm-d event interoperability", "classification": "inspiration", "disposition": "follow_on_open", "reason": "FAK issues #3888 and #6082 already own the bounded feature and matched benchmark. Reuse the protocol idea without yielding fak-native cache ownership." }, { "candidate": "pluginized endpoint scoring, flow control, P/D topology, and workload-variant autoscaling", "source": [ "docs/api-reference/endpointpickerconfig.md", "docs/architecture/advanced/autoscaling/", "docs/architecture/advanced/disaggregation/", "guides/" ], "alternative": "Implement a second Kubernetes inference control plane inside fak", "classification": "interoperation", "disposition": "reject_duplicate", "reason": "These are llm-d control-plane responsibilities. FAK should expose clean signals and compatibility seams, not port the scheduler, autoscaler, deployment guides, or Gateway API stack." } ], "issue_tracking": { "study_issue": "https://github.com/anthony-chaudhary/fak/issues/8988", "parent_issue": "https://github.com/anthony-chaudhary/fak/issues/8936", "shipped": [ "#8016" ], "open_follow_ons": [ "#8017", "#6082", "#3888" ], "new_filed_borrows": [], "reason": "Every surviving candidate is already shipped or durably owned. No duplicate issue was filed." } } }