feat(cdc): app-level Cassandra+Neo4j change events to Kafka (->S3) on generate

infra/airflow/scripts/generate_cassandra_telemetry_data.py

@@ -9,6 +9,13 @@ import random
 from datetime import datetime, timedelta
 import uuid
 import sys
+import os
+sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+try:
+    from cdc_emit import emit_changes
+except Exception:
+    def emit_changes(*a, **k):
+        return 0
 
 # Database connection details
 DB_HOST = "10.0.21.51"
@@ -57,12 +64,25 @@ def main():
     )
 
     total_generated = 0
+    cdc_buffer = []
+    CDC_TOPIC = "cassandra_telemetry.cdc.device_metrics"
     for i in range(TARGET_ROWS):
-        session.execute(insert_stmt, generate_fake_device_metric())
+        row = generate_fake_device_metric()
+        session.execute(insert_stmt, row)
         total_generated += 1
+        cdc_buffer.append({
+            "device_id": row[0], "metric_ts": row[1], "metric_type": row[2],
+            "metric_value": row[3], "payload": row[4],
+        })
+        if len(cdc_buffer) >= 1000:
+            emit_changes(CDC_TOPIC, "telemetry", "device_metrics", cdc_buffer)
+            cdc_buffer = []
         if total_generated % 1000 == 0:
             print(f"Generated {total_generated} rows...")
 
+    if cdc_buffer:
+        emit_changes(CDC_TOPIC, "telemetry", "device_metrics", cdc_buffer)
+
     session.shutdown()
     cluster.shutdown()