fix: Dashboard Total Traces count inflated by CIDX vector index files (v9.3.89)

_get_langfuse_folder_stats() counted all *.json files under langfuse_* folders
including .code-indexer/ vector store files (88K+), inflating the trace count
~22x (showed 92K instead of ~4K). Fixed by excluding .code-indexer from path.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: jsbattig

CHANGELOG.md

@@ -5,6 +5,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## v9.3.89
+
+### Bug Fixes
+
+- fix: Dashboard "Total Traces" count inflated by CIDX vector index files. The _get_langfuse_folder_stats() method used folder.glob("**/*.json") which counted .code-indexer/ vector store JSON files (88K+) alongside actual trace files (4K), inflating the reported count ~22x. Fixed by excluding files with .code-indexer in their path components.
+
 ## v9.3.88
 
 ### Enhancements

README.md

@@ -2,7 +2,7 @@
 
 AI-powered semantic code search for your codebase. Find code by meaning, not just keywords.
 
-**Version 9.3.88** - [Changelog](CHANGELOG.md) | [Migration Guide](docs/migration-to-v8.md) | [Architecture](docs/architecture.md)
+**Version 9.3.89** - [Changelog](CHANGELOG.md) | [Migration Guide](docs/migration-to-v8.md) | [Architecture](docs/architecture.md)
 
 ## Quick Navigation
 

src/code_indexer/__init__.py

@@ -6,5 +6,5 @@
 HNSW graph indexing (O(log N) complexity).
 """
 
-__version__ = "9.3.88"
+__version__ = "9.3.89"
 __author__ = "Seba Battig"

src/code_indexer/server/services/dashboard_service.py

@@ -605,9 +605,9 @@ def _get_langfuse_folder_stats(self, server_dir: Path) -> dict:
             if folder.is_dir() and folder.name.startswith("langfuse_"):
                 user_folders += 1
 
-                # Count JSON files and sizes
+                # Count JSON files and sizes (exclude CIDX index files)
                 for trace_file in folder.glob("**/*.json"):
-                    if trace_file.is_file():
+                    if trace_file.is_file() and ".code-indexer" not in trace_file.parts:
                         total_traces += 1
                         total_size_bytes += trace_file.stat().st_size
 

tests/unit/server/test_langfuse_dashboard.py

@@ -264,6 +264,33 @@ def test_count_trace_files(self):
             # Size should be > 0 (small files are rounded to 0.0, check >= 0)
             assert stats["total_size_mb"] >= 0.0
 
+    def test_code_indexer_files_excluded_from_count(self):
+        """JSON files under .code-indexer/ dirs must not count as traces."""
+        from code_indexer.server.services.dashboard_service import DashboardService
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            server_dir = Path(tmpdir)
+            golden_repos = server_dir / "data" / "golden-repos"
+            golden_repos.mkdir(parents=True)
+
+            user_folder = golden_repos / "langfuse_user1"
+            user_folder.mkdir()
+
+            # Real trace files
+            (user_folder / "trace1.json").write_text('{"id": "1"}')
+            (user_folder / "trace2.json").write_text('{"id": "2"}')
+
+            # CIDX index files — must NOT be counted
+            cidx_dir = user_folder / ".code-indexer" / "index" / "langfuse_traces"
+            cidx_dir.mkdir(parents=True)
+            (cidx_dir / "vector_0001.json").write_text('{"vec": [0.1, 0.2]}')
+            (cidx_dir / "vector_0002.json").write_text('{"vec": [0.3, 0.4]}')
+
+            service = DashboardService()
+            stats = service._get_langfuse_folder_stats(server_dir)
+            assert stats["user_folders"] == 1
+            assert stats["total_traces"] == 2  # only real traces, not CIDX files
+
     def test_calculate_storage_size(self):
         """Should calculate total storage size in MB."""
         from code_indexer.server.services.dashboard_service import DashboardService