feat(simple_metrics): add threshold-based alerting via send_notification

Compares each computed metric against optional per-model thresholds
from model_config.simple_metrics_thresholds. Convention:
- {metric}_max: breach when value > threshold (rmse, mse, mae)
- {metric}_min: breach when value < threshold (r2)

Fires WARNING notification on breach. Missing thresholds = no alerting.
Schema designed to be extensible for Card 2 (Drift) thresholds.

SIENTIAPDE-1986

laborious/activities/model_metrics.py

@@ -429,4 +429,34 @@ class ModelMetrics(SientiaMonitoring):
 
         self._debug_dataframe(f'Simple metrics dataframe: Size {data.shape}', data, metadata)
 
+        # Threshold alerting (optional — no crash when absent)
+        thresholds = input_data.get('thresholds')
+        if thresholds:
+            # Convention: _max thresholds breach when value > threshold,
+            #             _min thresholds breach when value < threshold.
+            for row in output_data:
+                metric_name = row['metric']
+                value = row['value']
+                max_key = f'{metric_name}_max'
+                min_key = f'{metric_name}_min'
+
+                breach_msg = None
+                if max_key in thresholds and value > thresholds[max_key]:
+                    breach_msg = f'{metric_name}={value} exceeds {max_key}={thresholds[max_key]}'
+                elif min_key in thresholds and value < thresholds[min_key]:
+                    breach_msg = f'{metric_name}={value} below {min_key}={thresholds[min_key]}'
+
+                if breach_msg:
+                    self.warning(
+                        f'Threshold breach for model {model_id}: {breach_msg}',
+                        metadata,
+                    )
+                    self.send_notification(
+                        metadata=metadata,
+                        notification_id='SIMPLE_METRICS_THRESHOLD_BREACH',
+                        message=f'Threshold breach for model {model_id}: {breach_msg}',
+                        block='model_metrics',
+                        level=NotificationLevel.WARNING,
+                    )
+
         return data.to_dict(orient='records')