feat: turn profiling k8s jobs into sample DGDR requests (#3864)

Signed-off-by: Hannah Zhang <hannahz@nvidia.com>
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
Signed-off-by: Hongkuan Zhou <tedzhouhk@gmail.com>
Co-authored-by: hongkuanz <hongkuanz@nvidia.com>
Co-authored-by: Hongkuan Zhou <tedzhouhk@gmail.com>

tests/planner/README.md

@@ -23,7 +23,7 @@ Use the pre-configured test deployment with sample profiling data, we provide th
 
 ### Option B: Use Your Own Profiling Results
 
-1. Run pre-deployment profiling for your specific setup. See the [pre-deployment profiling documentation](../../docs/benchmarks/pre_deployment_profiling.md) for detailed instructions.
+1. Run pre-deployment profiling for your specific setup. See the [pre-deployment profiling documentation](../../docs/benchmarks/sla_driven_profiling.md) for detailed instructions.
 
 ## Interpolator Testing
 

tests/profiler/test_profile_sla_aiconfigurator.py

@@ -27,6 +27,8 @@ class TestProfileSlaAiconfigurator:
     def trtllm_args(self):
         class Args:
             def __init__(self):
+                self.model = ""
+                self.dgd_image = ""
                 self.backend = "trtllm"
                 self.config = "components/backends/trtllm/deploy/disagg.yaml"
                 self.output_dir = "/tmp/test_profiling_results"

tests/profiler/test_profile_sla_dryrun.py

@@ -49,6 +49,8 @@ class TestProfileSLADryRun:
                 self.config = "components/backends/vllm/deploy/disagg.yaml"
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
+                self.model = ""
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 1
                 self.max_num_gpus_per_engine = 8
                 self.skip_existing_results = False
@@ -83,6 +85,8 @@ class TestProfileSLADryRun:
                 self.config = "components/backends/sglang/deploy/disagg.yaml"
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
+                self.model = ""
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 1
                 self.max_num_gpus_per_engine = 8
                 self.skip_existing_results = False
@@ -131,6 +135,8 @@ class TestProfileSLADryRun:
                 self.config = "components/backends/trtllm/deploy/disagg.yaml"
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
+                self.model = ""
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 1
                 self.max_num_gpus_per_engine = 8
                 self.skip_existing_results = False
@@ -172,6 +178,8 @@ class TestProfileSLADryRun:
                 self.config = "recipes/deepseek-r1/sglang/disagg-16gpu/deploy.yaml"
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
+                self.model = ""
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 8
                 self.max_num_gpus_per_engine = 32
                 self.skip_existing_results = False
@@ -233,6 +241,7 @@ class TestProfileSLADryRun:
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
                 self.model = "deepseek-ai/DeepSeek-R1-Distill-Llama-8B"  # Specify model for autogen
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 0  # Will be auto-generated
                 self.max_num_gpus_per_engine = 0  # Will be auto-generated
                 self.skip_existing_results = False
@@ -294,6 +303,7 @@ class TestProfileSLADryRun:
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
                 self.model = "deepseek-ai/DeepSeek-R1-Distill-Llama-8B"  # Specify model for autogen
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 0  # Will be auto-generated
                 self.max_num_gpus_per_engine = 0  # Will be auto-generated
                 self.skip_existing_results = False
@@ -355,6 +365,7 @@ class TestProfileSLADryRun:
                 self.output_dir = "/tmp/test_profiling_results"
                 self.namespace = "test-namespace"
                 self.model = "deepseek-ai/DeepSeek-R1-Distill-Llama-8B"  # Specify model for autogen
+                self.dgd_image = ""
                 self.min_num_gpus_per_engine = 0  # Will be auto-generated
                 self.max_num_gpus_per_engine = 0  # Will be auto-generated
                 self.skip_existing_results = False