Fix safely of access ctx in exception handlers (#855)

Author: wu-clan

backend/common/log.py

@@ -6,7 +6,6 @@
 
 from loguru import logger
 
-from backend.common.context import ctx
 from backend.core.conf import settings
 from backend.core.path_conf import LOG_DIR
 from backend.utils.timezone import timezone
@@ -78,12 +77,8 @@ def setup_logging() -> None:
 
     # request_id 过滤器
     def request_id_filter(record: logging.LogRecord) -> logging.LogRecord:
-        if ctx.exists():
-            rid = get_request_trace_id()
-            record['request_id'] = rid[: settings.TRACE_ID_LOG_LENGTH]
-        else:
-            record['request_id'] = settings.TRACE_ID_LOG_DEFAULT_VALUE
-
+        rid = get_request_trace_id()
+        record['request_id'] = rid[: settings.TRACE_ID_LOG_LENGTH]
         return record
 
     # 配置 loguru 处理器

backend/utils/trace_id.py

@@ -4,4 +4,6 @@
 
 def get_request_trace_id() -> str:
     """从请求头中获取追踪 ID"""
-    return ctx.get(settings.TRACE_ID_REQUEST_HEADER_KEY, settings.TRACE_ID_LOG_DEFAULT_VALUE)
+    if ctx.exists():
+        return ctx.get(settings.TRACE_ID_REQUEST_HEADER_KEY, settings.TRACE_ID_LOG_DEFAULT_VALUE)
+    return settings.TRACE_ID_LOG_DEFAULT_VALUE