fix(anomaly-detection): Force using spans_eap and normal mode (#90536)

Author: shruthilayaj

src/sentry/search/eap/constants.py

@@ -92,8 +92,8 @@
 }
 
 # https://github.com/getsentry/snuba/blob/master/snuba/web/rpc/v1/endpoint_time_series.py
-# The RPC limits us to 2016 points per timeseries
-MAX_ROLLUP_POINTS = 2016
+# The RPC limits us to 2688 points per timeseries
+MAX_ROLLUP_POINTS = 2688
 # Copied from snuba, a number of total seconds
 VALID_GRANULARITIES = frozenset(
     {

src/sentry/seer/anomaly_detection/utils.py

@@ -219,6 +219,9 @@ def get_dataset_from_label(dataset_label: str):
     dataset = get_dataset(dataset_label)
     if dataset is None:
         raise ParseError(detail=f"dataset must be one of: {', '.join(DATASET_OPTIONS.keys())}")
+
+    if dataset == spans_eap:
+        return spans_rpc
     return dataset
 
 
@@ -290,7 +293,7 @@ def fetch_historical_data(
                 auto_fields=False,
                 use_aggregate_conditions=False,
             ),
-            sampling_mode="BEST_EFFORT",
+            sampling_mode="NORMAL",
         )
         return results
     else:

tests/sentry/seer/anomaly_detection/test_store_data.py

@@ -171,15 +171,19 @@ def test_anomaly_detection_fetch_historical_data_eap_spans(self):
             )
 
             result = fetch_historical_data(
-                self.organization, snuba_query, ["count()"], self.project
+                self.organization, snuba_query, ["count(span.duration)"], self.project
             )
             assert result
-            assert {"time": int(expected_ts_1), "count()": 1} in result.data.get("data")
-            assert {"time": int(expected_ts_2), "count()": 1} in result.data.get("data")
+            assert {"time": int(expected_ts_1), "count(span.duration)": 1} in result.data.get(
+                "data"
+            )
+            assert {"time": int(expected_ts_2), "count(span.duration)": 1} in result.data.get(
+                "data"
+            )
 
             formatted_result = format_historical_data(
                 data=result,
-                query_columns=["count()"],
+                query_columns=["count(span.duration)"],
                 dataset=spans_rpc,
                 organization=self.organization,
             )