feat(insights): Add division_if for span metrics (#72995)

Add division_if for span metrics so we can calculate rates such as slow
and frozen rates.

---------

Co-authored-by: Markus Hintersteiner <markus.hintersteiner@sentry.io>

src/sentry/search/events/datasets/spans_metrics.py

@@ -177,6 +177,30 @@ def function_converter(self) -> Mapping[str, fields.MetricsFunction]:
                     ),
                     default_result_type="integer",
                 ),
+                fields.MetricsFunction(
+                    "division_if",
+                    required_args=[
+                        fields.MetricArg(
+                            # the dividend, needs to be named column, otherwise the query builder won't be able to determine the correct target table
+                            "column",
+                            allow_custom_measurements=False,
+                        ),
+                        fields.MetricArg(
+                            "divisorColumn",
+                            allow_custom_measurements=False,
+                        ),
+                        fields.MetricArg(
+                            "if_col",
+                            allowed_columns=["release"],
+                        ),
+                        fields.SnQLStringArg(
+                            "if_val", unquote=True, unescape_quotes=True, optional_unquote=True
+                        ),
+                    ],
+                    snql_gauge=self._resolve_division_if,
+                    snql_distribution=self._resolve_division_if,
+                    default_result_type="percentage",
+                ),
                 fields.MetricsFunction(
                     "sum",
                     optional_args=[
@@ -1228,6 +1252,48 @@ def _resolve_avg_if(self, args, alias):
             alias,
         )
 
+    def _resolve_sum_if(
+        self,
+        metric_name: str,
+        if_col_name: str,
+        if_val: SelectType,
+        alias: str | None = None,
+    ) -> SelectType:
+        return Function(
+            "sumIf",
+            [
+                Column("value"),
+                Function(
+                    "and",
+                    [
+                        Function(
+                            "equals",
+                            [
+                                Column("metric_id"),
+                                self.resolve_metric(metric_name),
+                            ],
+                        ),
+                        Function(
+                            "equals",
+                            [self.builder.column(if_col_name), if_val],
+                        ),
+                    ],
+                ),
+            ],
+            alias,
+        )
+
+    def _resolve_division_if(
+        self,
+        args: Mapping[str, str | Column | SelectType],
+        alias: str,
+    ) -> SelectType:
+        return function_aliases.resolve_division(
+            self._resolve_sum_if(args["column"], args["if_col"], args["if_val"]),
+            self._resolve_sum_if(args["divisorColumn"], args["if_col"], args["if_val"]),
+            alias,
+        )
+
     def _resolve_avg_compare(self, args, alias):
         return function_aliases.resolve_avg_compare(self.builder.column, args, alias)
 

tests/snuba/api/endpoints/test_organization_events_span_metrics.py

@@ -121,6 +121,81 @@ def test_count_if(self):
         assert data[0][fieldRelease2] == 1
         assert meta["dataset"] == "spansMetrics"
 
+    def test_division_if(self):
+        self.store_span_metric(
+            {
+                "min": 1,
+                "max": 1,
+                "sum": 1,
+                "count": 1,
+                "last": 1,
+            },
+            entity="metrics_gauges",
+            metric="mobile.slow_frames",
+            timestamp=self.three_days_ago,
+            tags={"release": "1.0.0"},
+        )
+        self.store_span_metric(
+            {
+                "min": 1,
+                "max": 1,
+                "sum": 15,
+                "count": 15,
+                "last": 1,
+            },
+            entity="metrics_gauges",
+            metric="mobile.total_frames",
+            timestamp=self.three_days_ago,
+            tags={"release": "1.0.0"},
+        )
+        self.store_span_metric(
+            {
+                "min": 1,
+                "max": 1,
+                "sum": 2,
+                "count": 2,
+                "last": 1,
+            },
+            entity="metrics_gauges",
+            metric="mobile.frozen_frames",
+            timestamp=self.three_days_ago,
+            tags={"release": "2.0.0"},
+        )
+        self.store_span_metric(
+            {
+                "min": 1,
+                "max": 1,
+                "sum": 10,
+                "count": 10,
+                "last": 1,
+            },
+            entity="metrics_gauges",
+            metric="mobile.total_frames",
+            timestamp=self.three_days_ago,
+            tags={"release": "2.0.0"},
+        )
+
+        fieldRelease1 = "division_if(mobile.slow_frames,mobile.total_frames,release,1.0.0)"
+        fieldRelease2 = "division_if(mobile.frozen_frames,mobile.total_frames,release,2.0.0)"
+
+        response = self.do_request(
+            {
+                "field": [fieldRelease1, fieldRelease2],
+                "query": "",
+                "project": self.project.id,
+                "dataset": "spansMetrics",
+                "statsPeriod": "7d",
+            }
+        )
+        assert response.status_code == 200, response.content
+        data = response.data["data"]
+        meta = response.data["meta"]
+        assert len(data) == 1
+        assert data[0][fieldRelease1] == 1 / 15
+        assert data[0][fieldRelease2] == 2 / 10
+
+        assert meta["dataset"] == "spansMetrics"
+
     def test_count_unique(self):
         self.store_span_metric(
             1,