test(planner): isolate planner-family suites [DYN-2534] (#7723)

tests/profiler/test_helpers_profile_sla.py → components/src/dynamo/profiler/tests/unit/test_helpers_profile_sla.py

@@ -9,15 +9,18 @@ where needed.
 """
 
 import os
-import sys
 from pathlib import Path
 from unittest.mock import patch
 
 import pytest
 import yaml
 
-project_root = Path(__file__).parent.parent.parent
-sys.path.insert(0, str(project_root))
+pytestmark = [
+    pytest.mark.pre_merge,
+    pytest.mark.gpu_0,
+    pytest.mark.unit,
+    pytest.mark.planner,
+]
 
 try:
     from dynamo.planner.config.planner_config import (

tests/profiler/test_helpers_rapid.py → components/src/dynamo/profiler/tests/unit/test_helpers_rapid.py

@@ -9,15 +9,17 @@ the end-to-end test suite.
 """
 
 import copy
-import sys
-from pathlib import Path
 from unittest.mock import patch
 
 import pandas as pd
 import pytest
 
-project_root = Path(__file__).parent.parent.parent
-sys.path.insert(0, str(project_root))
+pytestmark = [
+    pytest.mark.pre_merge,
+    pytest.mark.gpu_0,
+    pytest.mark.unit,
+    pytest.mark.planner,
+]
 
 try:
     from dynamo.profiler.rapid import _run_default_sim, _run_naive_fallback

tests/profiler/test_helpers_thorough.py → components/src/dynamo/profiler/tests/unit/test_helpers_thorough.py

@@ -8,15 +8,17 @@ require live K8s deployments and are covered by the mocked end-to-end tests
 in test_profile_sla_dgdr.py.
 """
 
-import sys
-from pathlib import Path
 from unittest.mock import patch
 
 import pandas as pd
 import pytest
 
-project_root = Path(__file__).parent.parent.parent
-sys.path.insert(0, str(project_root))
+pytestmark = [
+    pytest.mark.pre_merge,
+    pytest.mark.gpu_0,
+    pytest.mark.unit,
+    pytest.mark.planner,
+]
 
 try:
     from dynamo.profiler.thorough import _pick_thorough_best_config

tests/profiler/test_planner_image_selection.py → components/src/dynamo/profiler/tests/unit/test_planner_image_selection.py

 # SPDX-FileCopyrightText: Copyright (c) 2025-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 # SPDX-License-Identifier: Apache-2.0
 
-import sys
-from pathlib import Path
-
 import pytest
 
-project_root = Path(__file__).parent.parent.parent
-sys.path.insert(0, str(project_root))
-sys.path.insert(0, str(project_root / "components" / "src"))
-
-pytestmark = [pytest.mark.pre_merge, pytest.mark.gpu_0, pytest.mark.unit]
+pytestmark = [
+    pytest.mark.pre_merge,
+    pytest.mark.gpu_0,
+    pytest.mark.unit,
+    pytest.mark.planner,
+]
 
 try:
     from dynamo.profiler.utils.config import update_image

components/src/dynamo/profiler/tests/test_profile_sla_auto_backend.py → components/src/dynamo/profiler/tests/unit/test_profile_sla_auto_backend.py

@@ -11,6 +11,7 @@ pytestmark = [
     pytest.mark.unit,
     pytest.mark.gpu_0,
     pytest.mark.pre_merge,
+    pytest.mark.planner,
     pytest.mark.parallel,
 ]
 

components/src/dynamo/profiler/tests/test_profiler_protocol.py → components/src/dynamo/profiler/tests/unit/test_profiler_protocol.py

@@ -24,6 +24,7 @@ pytestmark = [
     pytest.mark.unit,
     pytest.mark.gpu_0,
     pytest.mark.pre_merge,
+    pytest.mark.planner,
     pytest.mark.parallel,
 ]
 

container/compliance/Dockerfile.extract

@@ -41,7 +41,11 @@ RUN --mount=type=bind,from=target,target=/target \
     python3 /helpers/python_helper.py --root /target > /output/python.tsv 2>/output/python_err.txt ; \
     cat /output/dpkg_err.txt >&2 ; \
     cat /output/python_err.txt >&2 ; \
+    if [ -f /target/var/lib/dpkg/status ]; then \
         [ -s /output/dpkg.tsv ] || { echo "ERROR: dpkg extraction produced no output" >&2; exit 1; } ; \
+    else \
+        echo "⚠️  WARNING: dpkg status file not present; skipping OS package extraction" >&2 ; \
+    fi ; \
     [ -s /output/python.tsv ] || echo "⚠️  WARNING: python extraction produced no output" >&2
 
 FROM scratch

container/compliance/helpers/dpkg_helper.py

@@ -120,6 +120,12 @@ def parse_dpkg_status(status_path):
                 elif ":" in line:
                     key, _, val = line.partition(":")
                     current[key.strip()] = val.strip()
+    except FileNotFoundError:
+        print(
+            f"WARNING: No dpkg status file found: {status_path}",
+            file=sys.stderr,
+        )
+        return {}
     except (OSError, IOError):
         print(f"ERROR: Cannot read dpkg status file: {status_path}", file=sys.stderr)
         sys.exit(1)

container/compliance/resolve_base_image.py

@@ -32,7 +32,7 @@ def main() -> None:
     parser.add_argument(
         "--target",
         default="runtime",
-        choices=["runtime", "frontend"],
+        choices=["runtime", "frontend", "planner"],
         help="Build target (default: runtime)",
     )
     parser.add_argument(
@@ -74,6 +74,25 @@ def main() -> None:
         print(image)
         return
 
+    if args.target == "planner":
+        if args.framework != "dynamo":
+            print(
+                "ERROR: --target planner is only supported for --framework dynamo",
+                file=sys.stderr,
+            )
+            sys.exit(1)
+        planner_cfg = ctx.get("dynamo", {})
+        runtime_image = planner_cfg.get("planner_runtime_image")
+        runtime_image_tag = planner_cfg.get("planner_runtime_image_tag")
+        if not runtime_image or not runtime_image_tag:
+            print(
+                "ERROR: planner_runtime_image/planner_runtime_image_tag not found in context.yaml dynamo section",
+                file=sys.stderr,
+            )
+            sys.exit(1)
+        print(f"{runtime_image}:{runtime_image_tag}")
+        return
+
     # Runtime target
     if not args.cuda_version:
         print("ERROR: --cuda-version is required for runtime targets", file=sys.stderr)

container/templates/planner.Dockerfile

@@ -63,9 +63,8 @@ RUN --mount=type=bind,source=./container/deps/requirements.planner.txt,target=/t
         /opt/dynamo/wheelhouse/ai_dynamo*any.whl
 
 # Copy only the subset of the repository needed for planner/profiler service
-# startup and targeted planner/profiler unit tests.
+# startup and the component-local planner-family test suites.
 COPY --chmod=664 --chown=dynamo:0 pyproject.toml /workspace/pyproject.toml
-COPY --chmod=775 --chown=dynamo:0 tests /workspace/tests
 COPY --chmod=775 --chown=dynamo:0 components/src/dynamo/planner /workspace/components/src/dynamo/planner
 COPY --chmod=775 --chown=dynamo:0 components/src/dynamo/profiler /workspace/components/src/dynamo/profiler
 COPY --chmod=775 --chown=dynamo:0 components/src/dynamo/global_planner /workspace/components/src/dynamo/global_planner
@@ -75,7 +74,7 @@ COPY --chmod=775 --chown=dynamo:0 examples /workspace/examples
 FROM ${PLANNER_RUNTIME_IMAGE}:${PLANNER_RUNTIME_IMAGE_TAG} AS planner
 
 COPY --from=planner_builder /etc/group /etc/passwd /etc/
-COPY --from=planner_builder /bin/dash /usr/bin/sh
+COPY --from=planner_builder /bin/dash /bin/sh
 COPY --from=planner_builder /bin/uv /bin/uvx /usr/local/bin/
 COPY --chown=1000:0 --from=planner_builder /home/dynamo /home/dynamo
 COPY --chown=1000:0 --from=planner_builder /opt/dynamo/venv /opt/dynamo/venv
@@ -90,7 +89,7 @@ ENV DYNAMO_COMMIT_SHA=${DYNAMO_COMMIT_SHA} \
     VIRTUAL_ENV=/opt/dynamo/venv \
     LD_LIBRARY_PATH="/opt/dynamo/lib" \
     PATH="/opt/dynamo/venv/bin:/usr/local/bin/etcd:/usr/local/bin:/bin" \
-    PYTHONPATH="/workspace"
+    PYTHONPATH="/workspace/components/src:/workspace"
 
 WORKDIR /workspace
 USER dynamo

docs/components/planner/README.md

@@ -86,7 +86,7 @@ args:
   - --loadbased-adjustment-interval=5
 ```
 
-The planner will auto-discover the frontend metrics endpoint from the DGD. See [disagg_planner_load.yaml](https://github.com/ai-dynamo/dynamo/blob/main/tests/planner/scaling/disagg_planner_load.yaml) for a complete example.
+The planner will auto-discover the frontend metrics endpoint from the DGD. See [disagg_planner.yaml](https://github.com/ai-dynamo/dynamo/blob/main/examples/backends/vllm/deploy/disagg_planner.yaml) for a complete example.
 
 ### Manual DGD Deployment
 

docs/components/planner/global-planner.md

@@ -239,7 +239,7 @@ The planner inside each pool must be configured for `global-planner` mode so it
   "ttft": 2000,
   "prefill_engine_num_gpu": 2,
   "model_name": "${MODEL_NAME}",
-  "profile_results_dir": "/workspace/tests/planner/profiling_results/H200_TP1P_TP1D"
+  "profile_results_dir": "/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"
 }
 ```
 

docs/mocker/mocker.md

@@ -241,7 +241,7 @@ It also accepts older raw-data directories containing:
 ```bash
 python -m dynamo.mocker \
     --model-path nvidia/Llama-3.1-8B-Instruct-FP8 \
-    --planner-profile-data tests/planner/profiling_results/H200_TP1P_TP1D \
+    --planner-profile-data components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D \
     --speedup-ratio 1.0
 ```
 

examples/backends/mocker/deploy/agg.yaml

@@ -34,4 +34,4 @@ spec:
             - --speedup-ratio
             - "1.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
\ No newline at end of file

examples/backends/mocker/deploy/disagg.yaml

@@ -34,7 +34,7 @@ spec:
             - --speedup-ratio
             - "1.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
             - --disaggregation-mode
             - prefill
     decode:
@@ -58,6 +58,6 @@ spec:
             - --speedup-ratio
             - "1.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
             - --disaggregation-mode
             - decode

examples/global_planner/global-planner-gpu-budget.yaml

@@ -168,7 +168,7 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"disagg","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"itl":200,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"decode_engine_num_gpu":1,"model_name":"${MODEL_A}","profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"disagg","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"itl":200,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"decode_engine_num_gpu":1,"model_name":"${MODEL_A}","profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"}'
 ---
 # ── Model B: self-contained disagg serving DGD ──────────────────────────────
 apiVersion: nvidia.com/v1alpha1
@@ -263,4 +263,4 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"disagg","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"itl":200,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"decode_engine_num_gpu":1,"model_name":"${MODEL_B}","profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"disagg","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"itl":200,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"decode_engine_num_gpu":1,"model_name":"${MODEL_B}","profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"}'

examples/global_planner/global-planner-mocker-test.yaml

@@ -164,7 +164,7 @@ spec:
             - --speedup-ratio
             - "5.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
             - --is-prefill-worker
 
     Planner:
@@ -181,7 +181,7 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"prefill","throughput_metrics_source":"router","throughput_adjustment_interval":30,"ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"prefill","throughput_metrics_source":"router","throughput_adjustment_interval":30,"ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
 ---
 apiVersion: nvidia.com/v1alpha1
 kind: DynamoGraphDeployment
@@ -232,7 +232,7 @@ spec:
             - --speedup-ratio
             - "5.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
             - --is-prefill-worker
 
     Planner:
@@ -249,7 +249,7 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"prefill","throughput_metrics_source":"router","throughput_adjustment_interval":30,"ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"prefill","throughput_metrics_source":"router","throughput_adjustment_interval":30,"ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
 ---
 apiVersion: nvidia.com/v1alpha1
 kind: DynamoGraphDeployment
@@ -301,7 +301,7 @@ spec:
             - --speedup-ratio
             - "5.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
 
     Planner:
       componentType: planner
@@ -317,7 +317,7 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"decode","throughput_metrics_source":"router","throughput_adjustment_interval":30,"itl":200,"max_gpu_budget":-1,"decode_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"decode","throughput_metrics_source":"router","throughput_adjustment_interval":30,"itl":200,"max_gpu_budget":-1,"decode_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
 ---
 apiVersion: nvidia.com/v1alpha1
 kind: DynamoGraphDeployment
@@ -369,7 +369,7 @@ spec:
             - --speedup-ratio
             - "5.0"
             - --planner-profile-data
-            - /workspace/tests/planner/profiling_results/H200_TP1P_TP1D
+            - /workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D
 
     Planner:
       componentType: planner
@@ -385,4 +385,4 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"decode","throughput_metrics_source":"router","throughput_adjustment_interval":30,"itl":200,"max_gpu_budget":-1,"decode_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"mocker","mode":"decode","throughput_metrics_source":"router","throughput_adjustment_interval":30,"itl":200,"max_gpu_budget":-1,"decode_engine_num_gpu":1,"no_correction":true,"profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D","model_name":"nvidia/Llama-3.1-8B-Instruct-FP8"}'

examples/global_planner/global-planner-vllm-test.yaml

@@ -233,7 +233,7 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"prefill","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"model_name":"${MODEL_NAME}","profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"prefill","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":1,"model_name":"${MODEL_NAME}","profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"}'
 ---
 apiVersion: nvidia.com/v1alpha1
 kind: DynamoGraphDeployment
@@ -308,7 +308,7 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"prefill","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":2,"model_name":"${MODEL_NAME}","profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"prefill","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","ttft":2000,"max_gpu_budget":-1,"prefill_engine_num_gpu":2,"model_name":"${MODEL_NAME}","profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"}'
 ---
 apiVersion: nvidia.com/v1alpha1
 kind: DynamoGraphDeployment
@@ -383,4 +383,4 @@ spec:
             - dynamo.planner
           args:
             - --config
-            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"decode","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","itl":200,"max_gpu_budget":-1,"decode_engine_num_gpu":1,"model_name":"${MODEL_NAME}","profile_results_dir":"/workspace/tests/planner/profiling_results/H200_TP1P_TP1D"}'
+            - '{"environment":"global-planner","global_planner_namespace":"${K8S_NAMESPACE}-gp-ctrl","backend":"vllm","mode":"decode","enable_load_scaling":false,"enable_throughput_scaling":true,"throughput_metrics_source":"router","itl":200,"max_gpu_budget":-1,"decode_engine_num_gpu":1,"model_name":"${MODEL_NAME}","profile_results_dir":"/workspace/components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"}'

lib/bindings/python/tests/replay/replay_utils.py

@@ -353,7 +353,7 @@ def _run_aic_static_point(backend_name: str, isl: int, osl: int, batch_size: int
 def _planner_profile_data_dir_path() -> Path:
     return (
         Path(__file__).resolve().parents[5]
-        / "tests/planner/profiling_results/H200_TP1P_TP1D"
+        / "components/src/dynamo/planner/tests/data/profiling_results/H200_TP1P_TP1D"
     )
 
 

pyproject.toml

@@ -314,6 +314,11 @@ ignore_missing_imports = true
 module = ["dynamo.*.tests.*", "dynamo.*.tests"]
 ignore_errors = true
 
+[[tool.mypy.overrides]]
+# Manual planner helpers are operational scripts, not typed library surfaces.
+module = ["dynamo.planner.manual.*"]
+ignore_errors = true
+
 [[tool.mypy.overrides]]
 # Skip mypy analysis on backend framework internals.
 # ignore_missing_imports silences import-not-found only when the backend