⚡ Bolt: Optimize psutil process iteration to lazily fetch expensive attributes

.jules/bolt.md

@@ -0,0 +1,3 @@
+## 2024-05-15 - psutil process iteration performance
+**Learning:** Querying expensive attributes like `memory_info` and `status` for all processes in `psutil.process_iter` incurs significant overhead, especially on Windows.
+**Action:** Only query basic attributes like `pid` and `name` initially, and lazily fetch expensive attributes wrapped in try/except blocks only for the specific target processes.

desktop_services/excel_service.py

@@ -206,17 +206,30 @@ def get_running_instances(self) -> List[Dict[str, Any]]:
     def _get_excel_process_snapshots(self) -> Dict[int, Dict[str, Any]]:
         """Find Excel processes with one psutil scan for refresh performance."""
         snapshots: Dict[int, Dict[str, Any]] = {}
-        for proc in psutil.process_iter(["pid", "name", "memory_info", "status"]):
+        # OPTIMIZATION: Only fetch basic attributes initially to avoid expensive calls for all processes
+        for proc in psutil.process_iter(["pid", "name"]):
             try:
                 info = proc.info
                 process_name = info.get("name") or ""
                 if process_name.lower() != "excel.exe":
                     continue
                 pid = int(info["pid"])
+
+                # Lazily fetch expensive attributes only for target processes
+                try:
+                    memory_info = proc.memory_info()
+                except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess, AttributeError):
+                    memory_info = None
+
+                try:
+                    status = proc.status()
+                except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess, AttributeError):
+                    status = "running"
+
                 snapshots[pid] = {
                     "name": process_name,
-                    "memory_mb": self._memory_mb_from_info(info.get("memory_info")),
-                    "status": info.get("status") or "running",
+                    "memory_mb": self._memory_mb_from_info(memory_info),
+                    "status": status,
                 }
             except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess):
                 continue

desktop_services/operations_cockpit_service.py

@@ -852,19 +852,32 @@ def _default_processes() -> list[dict[str, Any]]:
         if psutil is None:
             return []
         processes = []
-        for proc in psutil.process_iter(["pid", "name", "memory_info", "status"]):
+        # OPTIMIZATION: Only fetch basic attributes initially to avoid expensive calls for all processes
+        for proc in psutil.process_iter(["pid", "name"]):
             try:
                 info = proc.info
             except (psutil.NoSuchProcess, psutil.AccessDenied):
                 continue
             if str(info.get("name") or "").casefold() != "excel.exe":
                 continue
+
+            # Lazily fetch expensive attributes only for target processes
+            try:
+                memory_info = proc.memory_info()
+            except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess, AttributeError):
+                memory_info = None
+
+            try:
+                status = proc.status()
+            except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess, AttributeError):
+                status = "running"
+
             processes.append(
                 {
                     "pid": info.get("pid"),
                     "name": info.get("name"),
-                    "memory_info": info.get("memory_info"),
-                    "status": info.get("status"),
+                    "memory_info": memory_info,
+                    "status": status,
                 }
             )
         return processes

tests/test_excel_desktop_service.py

@@ -352,7 +352,7 @@ def status(self):
             "is_responsive": True,
         }
     ]
-    assert requested_fields == [["pid", "name", "memory_info", "status"]]
+    assert requested_fields == [["pid", "name"]]
     assert title_lookup_pids == [{42}]