feat(etl): Hadoop->Trino movement DAG + movements registry + autonomous ETL-agents

- DAG hadoop_to_trino (deployed to Airflow) + worker hadoop_to_trino.py (on the
  Hadoop master) move HDFS historical_sales -> iceberg.hadoop.historical_sales_hdfs.
- api/movements.py: movement registry, Airflow trigger+watch, run tracking
  (state/duration/rows), endpoints /api/movements, /{id}/run, /runs.
- agent_ops.py: ETL-agent loop autonomously triggers movements on an interval
  and logs each run; /api/agent-ops/etl/toggle + etl status.

api/agent_ops.py

@@ -49,6 +49,10 @@ DML_AGENT = "data-custodian"
 
 _INTERVAL = float(os.getenv("AGENT_DML_INTERVAL_SECONDS", "45"))
 _MAX_ROWS = int(os.getenv("AGENT_DML_MAX_ROWS", "5"))
+# ETL-agent: autonomously trigger data movements on an interval.
+_ETL_INTERVAL = float(os.getenv("ETL_AGENT_INTERVAL_SECONDS", "300"))
+# Movements the ETL-agent cycles through autonomously (must exist in movements.py).
+_ETL_ROTATION = [m for m in os.getenv("ETL_AGENT_MOVEMENTS", "hadoop_to_trino,gen_postgres,gen_mysql,gen_mongodb").split(",") if m]
 # How many agent-created rows we keep around per source before favouring DELETE.
 _POOL_CAP = int(os.getenv("AGENT_DML_POOL_CAP", "400"))
 
@@ -64,6 +68,16 @@ _state: dict[str, Any] = {
     "by_op": {"insert": 0, "update": 0, "delete": 0},
 }
 
+_etl_state: dict[str, Any] = {
+    "enabled": os.getenv("ETL_AGENT_ENABLED", "1") not in ("0", "false", "False", ""),
+    "interval": _ETL_INTERVAL,
+    "rotation": _ETL_ROTATION,
+    "idx": 0,
+    "runs_total": 0,
+    "last": None,
+    "started": False,
+}
+
 # ── Realistic value domains ─────────────────────────────────────────────────
 REGIONS = ["NA", "EU", "EMEA", "APAC", "LATAM"]
 CHANNELS = ["online", "retail", "partner", "wholesale", "direct"]
@@ -291,11 +305,49 @@ async def agent_dml_loop() -> None:
         await asyncio.sleep(max(5.0, float(_state["interval"])))
 
 
+async def etl_agent_loop() -> None:
+    """Background loop: ETL-agents autonomously trigger data movements and log
+    each run (rows, duration, status) to the feed."""
+    _etl_state["started"] = True
+    await asyncio.sleep(30)  # let the platform settle
+    while True:
+        try:
+            if _etl_state["enabled"] and _etl_state["rotation"]:
+                from movements import trigger_and_watch, MOVEMENT_BY_ID
+
+                mid = _etl_state["rotation"][_etl_state["idx"] % len(_etl_state["rotation"])]
+                _etl_state["idx"] += 1
+                if mid in MOVEMENT_BY_ID:
+                    result = await trigger_and_watch(mid, autonomous=True)
+                    _etl_state["runs_total"] += 1
+                    _etl_state["last"] = {"movement_id": mid, "state": result.get("state"),
+                                          "rows": result.get("rows"), "duration_s": result.get("duration_s"),
+                                          "ts": datetime.now(timezone.utc).isoformat()}
+        except Exception as exc:
+            _etl_state["last"] = {"error": str(exc), "ts": datetime.now(timezone.utc).isoformat()}
+        await asyncio.sleep(max(30.0, float(_etl_state["interval"])))
+
+
 # ── Endpoints ────────────────────────────────────────────────────────────────
 @router.get("/status")
 async def status() -> JSONResponse:
     pools = {k: len(v) for k, v in _pools.items()}
-    return JSONResponse({"ok": True, "pools": pools, **_state})
+    return JSONResponse({"ok": True, "pools": pools, "etl": _etl_state, **_state})
+
+
+@router.post("/etl/toggle")
+async def etl_toggle(body: dict[str, Any] = Body(default={})) -> JSONResponse:
+    if "enabled" in body:
+        _etl_state["enabled"] = bool(body["enabled"])
+    else:
+        _etl_state["enabled"] = not _etl_state["enabled"]
+    if "interval" in body:
+        try:
+            _etl_state["interval"] = max(30.0, float(body["interval"]))
+        except (TypeError, ValueError):
+            pass
+    await _emit(f"[etl-agent] Autonomous ETL movements {'ENABLED' if _etl_state['enabled'] else 'PAUSED'} by operator", "warn")
+    return JSONResponse({"ok": True, "enabled": _etl_state["enabled"], "interval": _etl_state["interval"]})
 
 
 @router.post("/toggle")