diff --git a/src/openutm_verification/core/clients/air_traffic/blue_sky_client.py b/src/openutm_verification/core/clients/air_traffic/blue_sky_client.py
index c98007a..ed38af9 100644
--- a/src/openutm_verification/core/clients/air_traffic/blue_sky_client.py
+++ b/src/openutm_verification/core/clients/air_traffic/blue_sky_client.py
@@ -141,7 +141,91 @@ async def generate_bluesky_sim_air_traffic_data_with_sensor_latency_issues(
         duration: int | None = None,
     ) -> list[list[FlightObservationSchema]]:
         """This method generates"""
-        flight_observations = self.generate_bluesky_sim_air_traffic_data(config_path=config_path, duration=duration)
+
+
+        scn_path = config_path or self.settings.simulation_config_path
+        duration_s = int(duration or self.settings.simulation_duration_seconds or 30)
+
+        sensor_ids = self.settings.sensor_ids
+
+        try:
+            # create a list of UUIDs with at least one UUID if session_ids is empty
+            sensor_ids = [UUID(x) for x in sensor_ids] if sensor_ids else [uuid.uuid4()]
+        except ValueError as exc:
+            logger.error(f"Invalid sensor ID in configuration, it should be a valid UUID: {exc}")
+            raise
+        current_sensor_id = sensor_ids[0]
+
+        if not scn_path:
+            raise ValueError("No scenario path provided. Provide config_path or set settings.simulation_config_path.")
+
+        # ---- Init BlueSky headless ----
+        # detached=True prevents UI/event loop from blocking.
+        # Use a temporary directory for BlueSky working files to avoid polluting ~/bluesky
+        # BlueSky's pathfinder.init() auto-creates required subdirs (scenario, plugins, output, cache)
+        flight_observations: list[list[FlightObservationSchema]] = []
+        with tempfile.TemporaryDirectory(prefix="openutm-bluesky-") as tmp_dir:
+            cfg_path = os.path.join(tmp_dir, "settings.cfg")
+            bs.init(mode="sim", detached=True, workdir=tmp_dir, configfile=cfg_path)
+
+            # Route console output to stdout (useful for debugging stack commands)
+            bs.scr = ScreenDummy()
+            now = arrow.now()
+            logger.info(f"Initializing BlueSky (headless) and loading scenario: {relative_path(scn_path)} with duration {duration_s}s")
+
+            # ---- Load scenario ----
+            # BlueSky scenario files (like scenario/DEMO/bluesky_flight.scn) are typically loaded with IC.
+            # NOTE: Use absolute paths if relative paths cause issues inside Docker.
+            bs.stack.stack(f"IC {scn_path}")
+
+            # Ensure 1 Hz stepping; FF starts fast-time running mode, but we will still step manually.
+            # Some setups work fine with DT 1 and calling bs.sim.step().
+            bs.stack.stack("DT 1.0")
+
+            # ---- Sample data at 1 Hz for duration_s seconds ----
+            # Store per-aircraft series
+            results_by_acid: dict[str, list[FlightObservationSchema]] = {}
+
+            for t in range(1, duration_s + 1):
+                # Advance sim by one step (DT=1 sec)
+                bs.sim.step()
+                timestamp = now.shift(seconds=t)
+                timestamp_microseconds = int(timestamp.float_timestamp * 1_000_000)  # microseconds
+
+                # Snapshot traffic arrays
+                acids: list[str] = list(getattr(bs.traf, "id", []))
+                lats: list[float] = _tolist(getattr(bs.traf, "lat", []))
+                lons: list[float] = _tolist(getattr(bs.traf, "lon", []))
+                alts: list[float] = _tolist(getattr(bs.traf, "alt", []))
+
+                for i, acid in enumerate(acids):
+                    lat = float(lats[i])
+                    lon = float(lons[i])
+                    alt_m_or_ft = float(alts[i])
+
+                    # BlueSky typically uses meters internally for alt, but some scenarios use FL/ft inputs.
+                    # We store altitude_mm as "millimeters"; keep it consistent with your schema.
+                    # If alt is actually feet, you can convert here: alt_m = alt_ft * 0.3048
+                    altitude_mm = alt_m_or_ft * 1000.0
+                    metadata = {"sensor_id": current_sensor_id} if current_sensor_id else {}
+
+                    obs = FlightObservationSchema(
+                        lat_dd=lat,
+                        lon_dd=lon,
+                        altitude_mm=altitude_mm,
+                        traffic_source=0,
+                        source_type=0,
+                        icao_address=acid,
+                        timestamp=timestamp_microseconds,
+                        metadata=metadata,
+                    )
+                    results_by_acid.setdefault(acid, []).append(obs)
+
+                    logger.debug(f"{acid:>6} lat={lat:.6f} lon={lon:.6f} alt_mm={altitude_mm:.1f}")
+
+            # Convert dict -> list[list[FlightObservationSchema]] with stable ordering
+            flight_observations = [results_by_acid[acid] for acid in sorted(results_by_acid.keys())]
+
 
         # This method modifies the retrieved simulation data by changing the timestamp and adding latency to the observed dataset
         LATENCY_PROBABILITY = 0.1  # 10% chance to have latency issues