[CI] Revive 8-GPU trace upload in nightly test workflow (#18820)

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

.github/workflows/nightly-test-nvidia.yml

@@ -120,6 +120,25 @@ jobs:
           cd test
           python3 run_suite.py --hw cuda --suite nightly-8-gpu-common --nightly --timeout-per-file=18000 --continue-on-error --auto-partition-id=${{ matrix.partition }} --auto-partition-size=4
 
+      - name: Publish traces to storage repo
+        if: always()
+        continue-on-error: true
+        env:
+          GITHUB_TOKEN: ${{ secrets.GH_PAT_FOR_NIGHTLY_CI_DATA }}
+          GITHUB_RUN_ID: ${{ github.run_id }}
+          GITHUB_RUN_NUMBER: ${{ github.run_number }}
+        run: |
+          TRACE_ARGS=""
+          for dir in test/performance_profiles_*/; do
+            [ -d "$dir" ] && TRACE_ARGS="$TRACE_ARGS --traces-dir $dir"
+          done
+          if [ -n "$TRACE_ARGS" ]; then
+            python3 scripts/ci/utils/publish_traces.py $TRACE_ARGS
+            find test/performance_profiles_*/ -name '*.json.gz' -delete
+          else
+            echo "No trace directories found, skipping publish"
+          fi
+
       - name: Run test
         timeout-minutes: 30
         env:
@@ -201,6 +220,25 @@ jobs:
           cd test
           IS_BLACKWELL=1 python3 run_suite.py --hw cuda --suite nightly-8-gpu-common --nightly --timeout-per-file=12000 --continue-on-error --auto-partition-id=${{ matrix.partition }} --auto-partition-size=4
 
+      - name: Publish traces to storage repo
+        if: always()
+        continue-on-error: true
+        env:
+          GITHUB_TOKEN: ${{ secrets.GH_PAT_FOR_NIGHTLY_CI_DATA }}
+          GITHUB_RUN_ID: ${{ github.run_id }}
+          GITHUB_RUN_NUMBER: ${{ github.run_number }}
+        run: |
+          TRACE_ARGS=""
+          for dir in test/performance_profiles_*/; do
+            [ -d "$dir" ] && TRACE_ARGS="$TRACE_ARGS --traces-dir $dir"
+          done
+          if [ -n "$TRACE_ARGS" ]; then
+            python3 scripts/ci/utils/publish_traces.py $TRACE_ARGS
+            find test/performance_profiles_*/ -name '*.json.gz' -delete
+          else
+            echo "No trace directories found, skipping publish"
+          fi
+
       - name: Collect performance metrics
         if: always()
         run: |

scripts/ci/utils/publish_traces.py

@@ -331,7 +331,20 @@ def copy_trace_files(source_dir, target_base_path):
 
 
 def publish_traces(traces_dir, run_id, run_number):
-    """Publish traces to GitHub repository in a single commit"""
+    """Publish traces from a single directory to GitHub repository in a single commit"""
+    target_base_path = f"traces/{run_id}"
+    files_to_upload = copy_trace_files(traces_dir, target_base_path)
+
+    if not files_to_upload:
+        print("No trace files found to upload")
+        return
+
+    print(f"Found {len(files_to_upload)} files to upload")
+    publish_traces_from_files(files_to_upload, run_id, run_number)
+
+
+def publish_traces_from_files(files_to_upload, run_id, run_number):
+    """Publish pre-collected trace files to GitHub repository in a single commit"""
     # Get environment variables
     token = os.getenv("GITHUB_TOKEN")
     if not token:
@@ -342,16 +355,6 @@ def publish_traces(traces_dir, run_id, run_number):
     repo_owner = "sglang-bot"
     repo_name = "sglang-ci-data"
     branch = "main"
-    target_base_path = f"traces/{run_id}"
-
-    # Copy trace files
-    files_to_upload = copy_trace_files(traces_dir, target_base_path)
-
-    if not files_to_upload:
-        print("No trace files found to upload")
-        return
-
-    print(f"Found {len(files_to_upload)} files to upload")
 
     # Verify token permissions before proceeding
     permission_check = verify_token_permissions(repo_owner, repo_name, token)
@@ -475,8 +478,10 @@ def main():
     parser.add_argument(
         "--traces-dir",
         type=str,
+        action="append",
+        dest="traces_dirs",
         required=True,
-        help="Traces directory to publish",
+        help="Traces directory to publish (can be specified multiple times)",
     )
     args = parser.parse_args()
 
@@ -490,12 +495,22 @@ def main():
         )
         sys.exit(1)
 
-    # Use traces directory
+    # Collect trace files from all directories
-    traces_dir = args.traces_dir
+    target_base_path = f"traces/{run_id}"
-    print(f"Processing traces from directory: {traces_dir}")
+    all_files = []
+    for traces_dir in args.traces_dirs:
+        print(f"Processing traces from directory: {traces_dir}")
+        files = copy_trace_files(traces_dir, target_base_path)
+        all_files.extend(files)
 
-    # Publish traces
+    if not all_files:
-    publish_traces(traces_dir, run_id, run_number)
+        print("No trace files found to upload across all directories")
+        return
+
+    print(f"Found {len(all_files)} total files to upload")
+
+    # Publish all collected traces in a single commit
+    publish_traces_from_files(all_files, run_id, run_number)
 
 
 if __name__ == "__main__":