v1.8.x MLFlow registry and live GPU metrics

.eslintrc.cjs

@@ -44,6 +44,19 @@ module.exports = {
             rules: {
                 'property-test-rules/no-hardcoded-numruns': 'error'
             }
+        },
+        {
+            // Module-header convention (docs/architecture/module-header-convention.md).
+            // WARNING during the consolidation program (Waves 1-8); promoted to
+            // 'error' in Wave 9 once every pattern-participating module conforms.
+            // Scope expands wave by wave; starts at the shared MCP server libs
+            // (Wave 2 factory lands here). Test files are excluded.
+            files: ['servers/lib/**/*.js'],
+            excludedFiles: ['**/*.test.js'],
+            plugins: ['property-test-rules'],
+            rules: {
+                'property-test-rules/require-module-header': 'warn'
+            }
         }
     ],
     ignorePatterns: [

.github/workflows/ci.yml

@@ -52,6 +52,14 @@ jobs:
         if: matrix.node-version == '24.x'
         run: node scripts/validate-servers.js
 
+      # MCP-server unit suites (servers/*/test/*.test.js). These use a custom
+      # node:assert harness (not mocha) and were previously orphaned from CI.
+      # Now a blocking gate — the runner (scripts/run-server-tests.js) discovers
+      # every suite by glob and exits non-zero if any fail.
+      - name: MCP server unit tests
+        if: matrix.node-version == '24.x'
+        run: npm run test:servers:unit
+
       - name: Python tests
         if: matrix.node-version == '24.x'
         run: |
@@ -74,6 +82,10 @@ jobs:
         if: matrix.node-version == '24.x'
         run: node scripts/schema-template-coverage.js --strict
 
+      - name: Validate serve-engine manifests
+        if: matrix.node-version == '24.x'
+        run: node scripts/validate-serve-manifests.js
+
   integration:
     name: Integration
     runs-on: ubuntu-latest

.gitignore

@@ -43,3 +43,9 @@ infra/bootstrap-modules/.hyperpod-helm-cache/
 docs/dev/*-pr-summary.md
 docs/dev/*-pr-description.md
 docs/dev/v*-release-notes.md
+
+# MLflow local tracking store (created when logging without a remote MLFLOW_TRACKING_URI)
+mlruns/
+mlflow.db
+mlflow.db-journal
+mlflow.db-wal

.husky/pre-commit

@@ -12,10 +12,19 @@ fi
 # 1. Lint staged JS files (fast — only changed files)
 npx lint-staged
 
-# 2. Python syntax check on staged .py files in templates/do/lib/python/
+# 2. Python syntax check on staged .py files.
+#    Skips EJS templates (files containing <% %> tags, e.g.
+#    templates/deploy_notebook_generator.py and templates/code/*.py): those are
+#    only valid Python AFTER the generator renders them, so ast.parse on the raw
+#    template is a false positive. Real runtime Python (templates/do/lib/python/,
+#    scripts/, servers/) has no EJS tags and is still checked.
 STAGED_PY=$(git diff --cached --name-only --diff-filter=ACM | grep '\.py$' || true)
 if [ -n "$STAGED_PY" ]; then
     for pyfile in $STAGED_PY; do
+        # Skip EJS templates — they are not raw Python until rendered.
+        if grep -q '<%' "$pyfile" 2>/dev/null; then
+            continue
+        fi
         if ! python3 -c "import ast; ast.parse(open('$pyfile').read())" 2>/dev/null; then
             echo "❌ Python syntax error in: $pyfile"
             exit 1

bin/cli.js

@@ -193,7 +193,7 @@ program.configureHelp({
             '  ./do/deploy && ./do/test',
             '',
             '  # 4. Benchmark performance',
-            '  ./do/benchmark --workload multi_turn_chat',
+            '  ./do/benchmark --workload multi-turn-chat',
             '',
             helper.styleTitle('All generation options:'),
             ''
@@ -243,6 +243,7 @@ program
     .option('--non-interactive', 'Run without prompts (requires --profile and --region)')
     .option('--name <name>', 'Bootstrap profile name (default: "default")')
     .option('--with <modules>', 'Comma-separated modules to provision (non-interactive; default: core,registry)')
+    .option('--module <name>', 'Redeploy only the named module\'s CDK stack (for update)')
     .option('--force', 'Force removal without confirmation')
     .option('--dry-run', 'Preview module changes (stacks, resources, profile updates) without provisioning or destroying anything')
     .option('--verify', 'Verify resources exist (for status)')

config/parameter-schema-v2.json

@@ -57,15 +57,15 @@
                     "transformers-tensorrt-llm",
                     "transformers-lmi",
                     "transformers-djl",
+                    "transformers-llama-cpp",
                     "triton-fil",
                     "triton-onnxruntime",
                     "triton-tensorflow",
                     "triton-pytorch",
                     "triton-vllm",
                     "triton-tensorrtllm",
                     "triton-python",
-                    "diffusors-vllm-omni",
-                    "marketplace"
+                    "diffusors-vllm-omni"
                 ]
             },
             "phase": "model",
@@ -91,7 +91,7 @@
         },
         "modelName": {
             "type": "string",
-            "description": "Model identifier (hf-org/model, s3://..., registry://..., marketplace://...)",
+            "description": "Model identifier (hf-org/model, s3://..., registry://...)",
             "cliFlag": "--model-name",
             "cliArgName": "name",
             "envVar": "ML_MODEL_NAME",
@@ -127,7 +127,7 @@
         },
         "deploymentTarget": {
             "type": "enum",
-            "description": "Deployment target (managed-inference, async-inference, batch-transform, hyperpod-eks)",
+            "description": "Deployment target (managed-inference, async-inference, batch-transform, hyperpod-eks, eks)",
             "cliFlag": "--deployment-target",
             "cliArgName": "target",
             "envVar": "ML_DEPLOYMENT_TARGET",
@@ -140,7 +140,8 @@
                     "realtime-inference",
                     "async-inference",
                     "batch-transform",
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ]
             },
             "phase": "infrastructure",
@@ -193,7 +194,7 @@
             "prompt": {
                 "message": "Instance type?",
                 "type": "input",
-                "when": "deploymentTarget !== 'hyperpod-eks'"
+                "when": "deploymentTarget !== 'hyperpod-eks' && deploymentTarget !== 'eks'"
             },
             "deprecated": false,
             "since": "0.1.0"
@@ -317,7 +318,11 @@
             "group": "lora",
             "appliesTo": {
                 "deploymentTargets": [
-                    "managed-inference"
+                    "managed-inference",
+                    "hyperpod-eks",
+                    "async-inference",
+                    "batch-transform",
+                    "eks"
                 ],
                 "architectures": [
                     "transformers"
@@ -356,7 +361,11 @@
             "group": "lora",
             "appliesTo": {
                 "deploymentTargets": [
-                    "managed-inference"
+                    "managed-inference",
+                    "hyperpod-eks",
+                    "async-inference",
+                    "batch-transform",
+                    "eks"
                 ],
                 "architectures": [
                     "transformers"
@@ -395,7 +404,11 @@
             "group": "lora",
             "appliesTo": {
                 "deploymentTargets": [
-                    "managed-inference"
+                    "managed-inference",
+                    "hyperpod-eks",
+                    "async-inference",
+                    "batch-transform",
+                    "eks"
                 ],
                 "architectures": [
                     "transformers"
@@ -1708,7 +1721,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1718,7 +1732,7 @@
             "prompt": {
                 "message": "HyperPod cluster name?",
                 "type": "input",
-                "when": "deploymentTarget === 'hyperpod-eks'"
+                "when": "deploymentTarget === 'hyperpod-eks' || deploymentTarget === 'eks'"
             },
             "deprecated": false,
             "since": "0.4.0"
@@ -1737,7 +1751,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1764,7 +1779,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1789,7 +1805,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1817,7 +1834,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1842,7 +1860,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1867,7 +1886,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1892,7 +1912,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1917,7 +1938,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -1942,7 +1964,8 @@
             "group": "hyperpod",
             "appliesTo": {
                 "deploymentTargets": [
-                    "hyperpod-eks"
+                    "hyperpod-eks",
+                    "eks"
                 ],
                 "architectures": [
                     "*"
@@ -2040,6 +2063,35 @@
             "deprecated": false,
             "since": "0.12.0"
         },
+        "engineFeature": {
+            "type": "string",
+            "description": "Enable an engine-specific feature declared in the selected engine's serve plugin (engine_features). Form NAME=VALUE, repeatable (e.g. radix_attention=true for SGLang, rolling_batch_backend=vllm for LMI). Resolved to the engine's real env var and validated against the manifest.",
+            "cliFlag": "--engine-feature",
+            "cliArgName": "NAME=VALUE",
+            "envVar": null,
+            "templateVar": null,
+            "configKey": "engineFeature",
+            "default": [],
+            "validation": {},
+            "phase": "features",
+            "group": "model",
+            "appliesTo": {
+                "deploymentTargets": [
+                    "*"
+                ],
+                "architectures": [
+                    "*"
+                ]
+            },
+            "widget": {
+                "section": "env-vars",
+                "inputType": "repeatable-kv"
+            },
+            "prompt": null,
+            "repeatable": true,
+            "deprecated": false,
+            "since": "1.8.0"
+        },
         "includeSample": {
             "type": "boolean",
             "description": "Include sample model code",