fix(billing-cost-management-mcp-server): correct the ri-performance coverage metrics example and default (#4834)

The `metrics` arg doc told callers to send '["HoursCoverage", "CostCoverage"]'.
GetReservationCoverage rejects those names: "Selected metrics are not valid.
Valid metrics are Hour, Unit, Cost. Provided metrics are HoursCoverage,
CostCoverage". The doc also said omitting `metrics` returns all metrics, but
Cost Explorer then returns only the hours section (CoverageHours).

The doc now gives a working example ('["Hour", "Cost"]'), lists the valid
values, says Unit is EC2 only ("Metric Unit is not supported for <service>"
for the other reservable services), and says the default is Hour only.

Tests: a contract test checks the metrics arg doc, and the two tests that
sent CoverageHours/CoverageCost as request metrics now send Hour/Cost.

Author: finklen-amazon

src/billing-cost-management-mcp-server/awslabs/billing_cost_management_mcp_server/tools/ri_performance_tools.py

@@ -137,7 +137,7 @@ async def ri_performance(
         start_date: Start date in YYYY-MM-DD format (inclusive). Defaults to 30 days ago if not provided.
         end_date: End date in YYYY-MM-DD format (exclusive). Defaults to today if not provided.
         granularity: Time granularity of the data (DAILY or MONTHLY). Defaults to DAILY. Ignored when group_by is provided (granularity and group_by are mutually exclusive).
-        metrics: List of metrics to retrieve for coverage as a JSON string (e.g., '["HoursCoverage", "CostCoverage"]'). Defaults to all metrics.
+        metrics: List of metrics to retrieve for coverage as a JSON string (e.g., '["Hour", "Cost"]'). Valid values are Hour, Unit and Cost; Unit is EC2 only. Defaults to Hour only.
         group_by: Optional grouping as a JSON string. For coverage: AZ, CACHE_ENGINE, DATABASE_ENGINE, DEPLOYMENT_OPTION, INSTANCE_TYPE, INSTANCE_TYPE_FAMILY, LINKED_ACCOUNT, PLATFORM, REGION, or TENANCY, depending on the service (see the tool description) (SERVICE is NOT valid — filter on it instead). For utilization: SUBSCRIPTION_ID only.
         filter: Optional filter as a JSON string (e.g., by service, region, or instance type). SERVICE defaults to EC2 when omitted; valid coverage SERVICE values are "Amazon Elastic Compute Cloud - Compute", "Amazon Relational Database Service", "Amazon ElastiCache", "Amazon MemoryDB", "Amazon Redshift", "Amazon OpenSearch Service", and "Amazon Elasticsearch Service" (use the full display name, not "EC2"/"RDS"). SubscriptionId values are numeric, not ARNs/UUIDs. A wide grouped/DAILY query may exceed the result size cap (ToolResultOversized) — scope it with a filter, a shorter window, or MONTHLY granularity.
         sort_by: Optional sorting configuration as a JSON string with key and direction (ASCENDING or DESCENDING).

src/billing-cost-management-mcp-server/tests/tools/test_ri_performance_tools.py

@@ -482,7 +482,7 @@ async def test_get_reservation_coverage_with_options(
             PAGINATION_COMPLETE,
         )
 
-        mock_metrics = ['CoverageHours', 'CoverageCost']
+        mock_metrics = ['Hour', 'Cost']
         mock_group_by = [{'Type': 'DIMENSION', 'Key': 'INSTANCE_TYPE'}]
         mock_filter = {
             'Dimensions': {'Key': 'SERVICE', 'Values': ['Amazon Elastic Compute Cloud - Compute']}
@@ -766,7 +766,7 @@ async def test_ri_performance_utilization(self, mock_context):
     async def test_ri_performance_with_all_params(self, mock_context):
         """Test ri_performance with all parameters."""
         # Setup
-        metrics = '["CoverageHours", "CoverageCost"]'
+        metrics = '["Hour", "Cost"]'
         group_by = '[{"Type": "DIMENSION", "Key": "INSTANCE_TYPE"}]'
         filter_expr = '{"Dimensions": {"Key": "SERVICE", "Values": ["Amazon Elastic Compute Cloud - Compute"]}}'
         sort_by = '{"Key": "CoverageHoursPercentage", "SortOrder": "DESCENDING"}'
@@ -1181,6 +1181,21 @@ async def test_description_states_subscription_id_format_and_error():
     assert 'SERVICE filter takes a single value' not in description
 
 
+@pytest.mark.asyncio
+async def test_metrics_arg_lists_the_values_cost_explorer_accepts():
+    """GetReservationCoverage takes Hour, Unit and Cost; other names are a ValidationException."""
+    tool = await ri_performance_server.get_tool('ri-performance')
+    assert tool is not None
+    metrics_doc = tool.parameters['properties']['metrics']['description']
+    assert '["Hour", "Cost"]' in metrics_doc
+    assert 'Valid values are Hour, Unit and Cost; Unit is EC2 only' in metrics_doc
+    # Without Metrics, Cost Explorer returns only CoverageHours.
+    assert 'Defaults to Hour only' in metrics_doc
+    assert 'Defaults to all metrics' not in metrics_doc
+    for wrong in ('HoursCoverage', 'CostCoverage', 'CoverageHours', 'CoverageCost'):
+        assert wrong not in metrics_doc
+
+
 @pytest.mark.asyncio
 async def test_ungrouped_utilization_ignores_max_items_and_next_token(mock_context):
     """GetReservationUtilization rejects any token without GroupBy, so none is sent."""