SIENTIAPDE-1645: Use absolute time (time.time()) for training start timestamps instead of monotonic time.

model_manager/activities/training.py

@@ -315,7 +315,7 @@ class Training(SientiaMonitoring):
         metadata: dict | None,
     ) -> TrainModelResult:
         labels = self._get_training_labels(train_params)
-        start_time = time.monotonic()
+        start_time = time.time()
         try:
             return self.data_manager_repository.prepare_training_data(
                 train_file_bytes=train_bytes,
@@ -353,7 +353,7 @@ class Training(SientiaMonitoring):
             metadata,
         )
 
-        start_time = time.monotonic()
+        start_time = time.time()
         try:
             wrapper.train(
                 train_data=train_data,