logging

kernel_ai/hooks.py

@@ -26,17 +26,28 @@ def add_headers(response):
 
         started_ns = getattr(g, "request_started_ns", None)
         duration_ns = (time.time_ns() - started_ns) if started_ns else None
+        source_ip = request.headers.get("X-Forwarded-For", request.remote_addr)
+        if source_ip and "," in source_ip:
+            source_ip = source_ip.split(",")[0].strip()
         _http_access_logger.info(
             "http_request",
             extra={
                 "event_data": {
+                    "@timestamp": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
+                    "event.category": "web",
+                    "event.type": "access",
                     "event.dataset": "kernel_ai.http",
                     "request.id": request_id,
+                    "source.ip": source_ip,
                     "http.request.method": request.method,
+                    "http.version": request.environ.get("SERVER_PROTOCOL", ""),
+                    "host.name": request.host,
                     "url.path": request.path,
+                    "url.query": (request.query_string or b"").decode("utf-8", errors="ignore"),
+                    "user_agent.original": request.headers.get("User-Agent"),
                     "http.response.status_code": response.status_code,
+                    "service.environment": current_app.config.get("ENV", "production"),
                     "event.duration": duration_ns,
-                    "network.client.ip": request.headers.get("X-Forwarded-For", request.remote_addr),
                 }
             },
         )

kernel_ai/http/common.py

@@ -1,6 +1,21 @@
 """Common HTTP helpers."""
 
-from flask import jsonify
+from flask import g, has_request_context, jsonify
+
+
+def build_error_payload(message, code, details=None):
+    """Build a stable API error envelope."""
+    payload = {
+        "error": str(message),
+        "code": str(code),
+    }
+    if has_request_context():
+        request_id = getattr(g, "request_id", None)
+        if request_id:
+            payload["request_id"] = request_id
+    if details is not None:
+        payload["details"] = details
+    return payload
 
 
 def api_json(producer, error_status=500, error_extra=None, exception_statuses=None):
@@ -9,12 +24,19 @@ def api_json(producer, error_status=500, error_extra=None, exception_statuses=No
         return jsonify(producer())
     except Exception as e:
         status = error_status
+        error_code = "internal_error"
         if exception_statuses:
             for exc_type, exc_status in exception_statuses:
                 if isinstance(e, exc_type):
                     status = exc_status
                     break
-        payload = {"error": str(e)}
+        if status == 400:
+            error_code = "bad_request"
+        elif status == 404:
+            error_code = "not_found"
+        elif status == 503:
+            error_code = "service_unavailable"
+        payload = build_error_payload(str(e), error_code)
         if error_extra:
             payload.update(error_extra)
         return jsonify(payload), status

kernel_ai/services/network.py

@@ -11,6 +11,7 @@
 
 import psutil
 
+from kernel_ai.logging_helpers import log_event
 from kernel_ai.services.infra_utils import resolve_binary
 
 logger = logging.getLogger(__name__)
@@ -63,7 +64,16 @@ def hex_to_ip(hex_str):
                         }
                     )
         return connections[:20]
-    except (OSError, ValueError, IndexError):
+    except (OSError, ValueError, IndexError) as exc:
+        log_event(
+            logger,
+            "DEBUG",
+            "Failed to parse /proc/net/tcp, using mock active connections",
+            event_dataset="kernel_ai.app",
+            component="services.network",
+            operation="get_active_connections",
+            event_data={"error": str(exc)},
+        )
         return get_mock_active_connections()
 
 
@@ -396,6 +406,15 @@ def get_traceroute_info(remote_ip, max_hops=8, traceroute_cache=None, cache_ttl_
         if target_ip.is_loopback or target_ip.is_unspecified:
             return {"remote_ip": remote_ip, "tool": None, "reached": False, "hop_count": 0, "hops": [], "note": "Local address, traceroute skipped"}
     except ValueError:
+        log_event(
+            logger,
+            "DEBUG",
+            "Invalid IP passed to traceroute info",
+            event_dataset="kernel_ai.app",
+            component="services.network",
+            operation="get_traceroute_info",
+            event_data={"remote_ip": remote_ip},
+        )
         raise ValueError("Invalid IP address")
 
     now = time.time()

kernel_ai/services/processes_runtime.py

@@ -12,6 +12,8 @@
 
 import psutil
 
+from kernel_ai.logging_helpers import log_event
+
 logger = logging.getLogger(__name__)
 
 
@@ -72,7 +74,15 @@ def get_processes_detailed_data() -> list[dict]:
         except (psutil.NoSuchProcess, psutil.AccessDenied, psutil.ZombieProcess):
             continue
         except (OSError, ValueError, TypeError, KeyError) as exc:
-            logger.debug("Skipping process in detailed scan due to unexpected data: %s", exc)
+            log_event(
+                logger,
+                "DEBUG",
+                "Skipping process in detailed scan due to unexpected data",
+                event_dataset="kernel_ai.app",
+                component="services.processes_runtime",
+                operation="get_processes_detailed_data",
+                event_data={"error": str(exc)},
+            )
             continue
     return processes
 
@@ -86,7 +96,15 @@ def _parse_meminfo_kb():
                 if len(parts) >= 2 and parts[1].isdigit():
                     out[parts[0].rstrip(":")] = int(parts[1])
     except OSError as exc:
-        logger.debug("Failed to read /proc/meminfo: %s", exc)
+        log_event(
+            logger,
+            "DEBUG",
+            "Failed to read /proc/meminfo",
+            event_dataset="kernel_ai.app",
+            component="services.processes_runtime",
+            operation="_parse_meminfo_kb",
+            event_data={"error": str(exc)},
+        )
     return out
 
 
@@ -286,7 +304,15 @@ def collect_processes_realtime():
                 }
             )
         except (psutil.Error, OSError, ValueError, TypeError, KeyError) as exc:
-            logger.debug("Skipping process in realtime collection: %s", exc)
+            log_event(
+                logger,
+                "DEBUG",
+                "Skipping process in realtime collection",
+                event_dataset="kernel_ai.app",
+                component="services.processes_runtime",
+                operation="collect_processes_realtime",
+                event_data={"error": str(exc)},
+            )
             continue
     syscall_nodes.sort(key=lambda x: x.get("syscall_pressure", 0), reverse=True)
     syscall_nodes = syscall_nodes[:14]
@@ -320,7 +346,15 @@ def collect_processes_realtime():
             if status:
                 bucket["states"][status] = bucket["states"].get(status, 0) + 1
     except (psutil.Error, OSError) as exc:
-        logger.debug("Failed to sample net connections in processes realtime: %s", exc)
+        log_event(
+            logger,
+            "DEBUG",
+            "Failed to sample net connections in processes realtime",
+            event_dataset="kernel_ai.app",
+            component="services.processes_runtime",
+            operation="collect_processes_realtime",
+            event_data={"error": str(exc)},
+        )
 
     network_tracing = []
     for _, row in network_nodes.items():
@@ -450,7 +484,15 @@ def _add_edge(src_pid, dst_pid, edge_type, weight):
         strip_rows, mem_summary = _build_memory_visual_rows(meminfo_kb, syscall_nodes, vm, swap)
         memory_visual = {"layout": "strips", "rows": strip_rows, "summary": mem_summary}
     except (psutil.Error, OSError, ValueError, TypeError) as exc:
-        logger.debug("Failed to build memory visual payload: %s", exc)
+        log_event(
+            logger,
+            "DEBUG",
+            "Failed to build memory visual payload",
+            event_dataset="kernel_ai.app",
+            component="services.processes_runtime",
+            operation="collect_processes_realtime",
+            event_data={"error": str(exc)},
+        )
         memory_visual = {
             "layout": "strips",
             "rows": [],