fix: Reduce watch service CPU usage by increasing reload interval

Changed watch_project_reload_interval default from 30s to 300s (5 min)
to reduce CPU usage caused by frequent watcher restarts.

The watch service was restarting every 30 seconds to reload the project
list, causing the watchfiles library to rebuild its internal state.
This was especially noticeable on systems with many files.

Impact:
- Reduces watcher restart frequency by 10x
- Should reduce idle CPU usage from 20-30% to 2-3% or less
- New projects picked up within 5 min instead of 30 sec
- Users can configure lower values if needed

Fixes #448

Co-authored-by: Paul Hernandez <phernandez@users.noreply.github.com>
Signed-off-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>

Author: github-actions[bot]

src/basic_memory/config.py

@@ -122,7 +122,9 @@ class BasicMemoryConfig(BaseSettings):
     )
 
     watch_project_reload_interval: int = Field(
-        default=30, description="Seconds between reloading project list in watch service", gt=0
+        default=300,
+        description="Seconds between reloading project list in watch service. Higher values reduce CPU usage by minimizing watcher restarts. Default 300s (5 min) balances efficiency with responsiveness to new projects.",
+        gt=0,
     )
 
     # update permalinks on move

tests/sync/test_watch_service_reload.py

@@ -85,8 +85,8 @@ def stop_after_one_call(*args):
         with patch.object(watch_service, "write_status", return_value=None):
             await watch_service.run()
 
-    # Should have slept for 30 seconds when no projects found
-    mock_sleep.assert_called_with(30)
+    # Should have slept for the configured reload interval when no projects found
+    mock_sleep.assert_called_with(config.watch_project_reload_interval)
 
 
 @pytest.mark.asyncio