diff --git a/policies/idle_kusto_cluster.sql b/policies/idle_kusto_cluster.sql
new file mode 100644
index 0000000..7803cec
--- /dev/null
+++ b/policies/idle_kusto_cluster.sql
@@ -0,0 +1,18 @@
+-- Real check: Azure Data Explorer cluster with zero ingestion and zero
+-- queries over the lookback window while its reserved instance count stays
+-- active and billed hourly. This is a genuine reserved-capacity cost pattern
+-- rather than a transient idle condition.
+SELECT
+    'azure' AS provider,
+    resource_id,
+    'Azure Data Explorer Cluster' AS service_name,
+    ROUND(instance_count * hourly_price_per_instance * 730, 2) AS billed_cost,
+    resource_name,
+    sku,
+    instance_count,
+    total_ingestion_mb,
+    total_queries,
+    lookback_days,
+    'cluster with 0 ingestion and 0 queries over ' || lookback_days || ' days (' || sku || ')' AS evidence_reason
+FROM fact_idle_kusto_cluster
+WHERE total_ingestion_mb = 0 AND total_queries = 0;
diff --git a/src/cloudcost/engine/runner.py b/src/cloudcost/engine/runner.py
index 6bfd6f1..276b687 100644
--- a/src/cloudcost/engine/runner.py
+++ b/src/cloudcost/engine/runner.py
@@ -61,6 +61,7 @@
     import cloudcost.sources.azure.idle_ai_search
     import cloudcost.sources.azure.idle_aml_compute_instance
     import cloudcost.sources.azure.idle_aml_compute_cluster
+    import cloudcost.sources.azure.idle_kusto_cluster
     import cloudcost.sources.oci.billing
     import cloudcost.sources.alibaba.billing
 except ImportError as e:
diff --git a/src/cloudcost/sources/azure/idle_kusto_cluster.py b/src/cloudcost/sources/azure/idle_kusto_cluster.py
new file mode 100644
index 0000000..47a659f
--- /dev/null
+++ b/src/cloudcost/sources/azure/idle_kusto_cluster.py
@@ -0,0 +1,201 @@
+import json
+import subprocess
+from datetime import datetime, timedelta, timezone
+from typing import Any
+
+import pyarrow as pa
+
+from cloudcost.core.registry import registry
+
+
+def _normalize_region(location: str) -> str:
+    return location.lower().replace(" ", "")
+
+
+def _fetch_hourly_price(sku_name: str, region: str) -> float:
+    region = _normalize_region(region)
+    filter_str = (
+        f"armRegionName eq '{region}' and armSkuName eq '{sku_name}' "
+        f"and priceType eq 'Consumption' and contains(productName, 'Azure Data Explorer') eq true"
+    )
+
+    try:
+        raw = subprocess.run(
+            [
+                "curl",
+                "-s",
+                "-G",
+                "https://prices.azure.com/api/retail/prices",
+                "--data-urlencode",
+                f"$filter={filter_str}",
+            ],
+            capture_output=True,
+            text=True,
+            check=True,
+            timeout=15,
+        ).stdout
+        data = json.loads(raw)
+        items = [
+            item
+            for item in data.get("Items", [])
+            if item.get("unitOfMeasure") == "1 Hour"
+            and item.get("skuName") == sku_name
+        ]
+        return float(items[0]["retailPrice"]) if items else 0.0
+    except Exception:
+        return 0.0
+
+
+# Real check: Azure Data Explorer clusters reserve fixed instance counts
+# and are billed hourly even when ingestion and query volume is near zero.
+# The check is based on live cluster inventory plus zero-ingestion and zero-
+# query totals over the lookback window.
+@registry.register_source("azure.idle_kusto_cluster")
+class AzureIdleKustoClusterSource:
+    def __init__(self, config: dict):
+        self.resource_group = config.get("resource_group")
+        self.lookback_days = config.get("lookback_days", 7)
+
+        if not self.resource_group:
+            raise ValueError("azure.idle_kusto_cluster requires 'resource_group' in config")
+
+    def extract(self, context: Any = None) -> pa.Table:
+        clusters_raw = subprocess.run(
+            [
+                "az",
+                "kusto",
+                "cluster",
+                "list",
+                "--resource-group",
+                self.resource_group,
+                "--query",
+                "[].{id:id,name:name,sku:sku,location:location}",
+                "-o",
+                "json",
+            ],
+            capture_output=True,
+            text=True,
+            check=True,
+        ).stdout
+        clusters = json.loads(clusters_raw)
+
+        end_time = datetime.now(timezone.utc)
+        start_time = end_time - timedelta(days=self.lookback_days)
+
+        rows = []
+        for cluster in clusters:
+            sku = cluster.get("sku") or {}
+            if isinstance(sku, dict):
+                sku_name = sku.get("name") or sku.get("skuName") or "unknown"
+                instance_count = sku.get("capacity") or 1
+            else:
+                sku_name = sku
+                instance_count = 1
+
+            if sku_name == "unknown":
+                continue
+
+            ingestion_raw = subprocess.run(
+                [
+                    "az",
+                    "monitor",
+                    "metrics",
+                    "list",
+                    "--resource",
+                    cluster["id"],
+                    "--metric",
+                    "IngestionVolumeMB",
+                    "--aggregation",
+                    "Total",
+                    "--interval",
+                    "PT1H",
+                    "--start-time",
+                    start_time.strftime("%Y-%m-%dT%H:%M:%SZ"),
+                    "--end-time",
+                    end_time.strftime("%Y-%m-%dT%H:%M:%SZ"),
+                ],
+                capture_output=True,
+                text=True,
+                check=True,
+            ).stdout
+            ingestion_parsed = json.loads(ingestion_raw)
+
+            total_ingestion_mb = 0.0
+            for timeseries in ingestion_parsed.get("value", []):
+                for series in timeseries.get("timeseries", []):
+                    for point in series.get("data", []):
+                        total_ingestion_mb += point.get("total") or 0.0
+
+            query_raw = subprocess.run(
+                [
+                    "az",
+                    "monitor",
+                    "metrics",
+                    "list",
+                    "--resource",
+                    cluster["id"],
+                    "--metric",
+                    "QueryCount",
+                    "--aggregation",
+                    "Total",
+                    "--interval",
+                    "PT1H",
+                    "--start-time",
+                    start_time.strftime("%Y-%m-%dT%H:%M:%SZ"),
+                    "--end-time",
+                    end_time.strftime("%Y-%m-%dT%H:%M:%SZ"),
+                ],
+                capture_output=True,
+                text=True,
+                check=True,
+            ).stdout
+            query_parsed = json.loads(query_raw)
+
+            total_queries = 0.0
+            for timeseries in query_parsed.get("value", []):
+                for series in timeseries.get("timeseries", []):
+                    for point in series.get("data", []):
+                        total_queries += point.get("total") or 0.0
+
+            rows.append(
+                {
+                    "resource_id": cluster["id"].lower(),
+                    "resource_name": cluster["name"],
+                    "sku": sku_name,
+                    "instance_count": instance_count,
+                    "total_ingestion_mb": total_ingestion_mb,
+                    "total_queries": total_queries,
+                    "hourly_price_per_instance": _fetch_hourly_price(
+                        sku_name,
+                        cluster.get("location", "eastus"),
+                    ),
+                    "lookback_days": self.lookback_days,
+                }
+            )
+
+        if not rows:
+            return pa.table(
+                {
+                    "resource_id": pa.array([], type=pa.string()),
+                    "resource_name": pa.array([], type=pa.string()),
+                    "sku": pa.array([], type=pa.string()),
+                    "instance_count": pa.array([], type=pa.int64()),
+                    "total_ingestion_mb": pa.array([], type=pa.float64()),
+                    "total_queries": pa.array([], type=pa.float64()),
+                    "hourly_price_per_instance": pa.array([], type=pa.float64()),
+                    "lookback_days": pa.array([], type=pa.int64()),
+                }
+            )
+
+        return pa.table(
+            {
+                "resource_id": [r["resource_id"] for r in rows],
+                "resource_name": [r["resource_name"] for r in rows],
+                "sku": [r["sku"] for r in rows],
+                "instance_count": [r["instance_count"] for r in rows],
+                "total_ingestion_mb": [r["total_ingestion_mb"] for r in rows],
+                "total_queries": [r["total_queries"] for r in rows],
+                "hourly_price_per_instance": [r["hourly_price_per_instance"] for r in rows],
+                "lookback_days": [r["lookback_days"] for r in rows],
+            }
+        )
diff --git a/tests/policies/test_idle_kusto_cluster.py b/tests/policies/test_idle_kusto_cluster.py
new file mode 100644
index 0000000..16c7ed4
--- /dev/null
+++ b/tests/policies/test_idle_kusto_cluster.py
@@ -0,0 +1,47 @@
+import pyarrow as pa
+
+
+def kusto_rows(*rows: tuple[str, str, int, float, float, float]) -> pa.Table:
+    return pa.table(
+        {
+            "resource_id": pa.array([r[0] for r in rows], type=pa.string()),
+            "resource_name": pa.array([f"kusto-{r[0]}" for r in rows], type=pa.string()),
+            "sku": pa.array([r[1] for r in rows], type=pa.string()),
+            "instance_count": pa.array([r[2] for r in rows], type=pa.int64()),
+            "total_ingestion_mb": pa.array([r[3] for r in rows], type=pa.float64()),
+            "total_queries": pa.array([r[4] for r in rows], type=pa.float64()),
+            "hourly_price_per_instance": pa.array([r[5] for r in rows], type=pa.float64()),
+            "lookback_days": pa.array([7 for _ in rows], type=pa.int64()),
+        }
+    )
+
+
+def test_kusto_cluster_with_zero_ingestion_and_queries_is_reported(run_policy) -> None:
+    findings = run_policy(
+        "idle_kusto_cluster",
+        {
+            "fact_idle_kusto_cluster": kusto_rows(
+                ("cluster1", "Standard_E2a_v4", 2, 0.0, 0.0, 0.12)
+            )
+        },
+    )
+
+    assert len(findings) == 1
+    finding = findings[0]
+    assert finding.provider == "azure"
+    assert finding.service_name == "Azure Data Explorer Cluster"
+    assert finding.estimated_impact == 175.2
+    assert finding.evidence["evidence_reason"] == "cluster with 0 ingestion and 0 queries over 7 days (Standard_E2a_v4)"
+
+
+def test_kusto_cluster_with_activity_is_not_reported(run_policy) -> None:
+    findings = run_policy(
+        "idle_kusto_cluster",
+        {
+            "fact_idle_kusto_cluster": kusto_rows(
+                ("cluster1", "Standard_E2a_v4", 2, 150.0, 12.0, 0.12)
+            )
+        },
+    )
+
+    assert findings == []