[processor/transform] Update is_monotonic and aggregation_temporality…

… accessor functions (open-telemetry#10197)

* Update accessor functions

* Update readme

* Update changelog

* Update changelog

CHANGELOG.md

@@ -27,6 +27,7 @@
 - `datadogexporter`: add error checks for datadog exporter (#9964)
 - `groupbyattrsprocessor`: copied aggregationtemporality when grouping metrics. (#9088)
 - `mongodbreceiver`: Fix issue where receiver startup could hang (#10111)
+- `transformprocessor`: Fix issue where metric.aggregation_temporality and metric.is_monotic were not actually gettable or settable (#10197)
 
 ## v0.51.0
 

processor/transformprocessor/README.md

@@ -15,6 +15,8 @@ in the OTLP protobuf definition. e.g., `status.code`, `attributes["http.method"]
 `resource.` or `instrumentation_library.`, it will reference those values.  For metrics, `name`, `description`, `unit`, `type`, `is_monotonic`, and `aggregation_temporality` are accessed via `metric.`
   - The name `instrumentation_library` within OpenTelemetry is currently under discussion and may be changed in the future.
   - Metric data types are `None`, `Gauge`, `Sum`, `Histogram`, `ExponentialHistogram`, and `Summary`
+  - `aggregation_temporality` is converted to and from the [protobuf's numeric definition](https://github.com/open-telemetry/opentelemetry-proto/blob/main/opentelemetry/proto/metrics/v1/metrics.proto#L291).  Interact with this field using 0, 1, or 2.
+  - Until the grammar can handle booleans, `is_monotic` is handled via strings the strings `"true"` and `"false"`.
 - Literals: Strings, ints, and floats can be referenced as literal values
 - Function invocations: Functions can be invoked with arguments matching the function's expected arguments
 - Where clause: Telemetry to modify can be filtered by appending `where a <op> b`, with `a` and `b` being any of the above.

processor/transformprocessor/internal/metrics/metrics.go

@@ -15,6 +15,7 @@
 package metrics // import "github.com/open-telemetry/opentelemetry-collector-contrib/processor/transformprocessor/internal/metrics"
 import (
 	"fmt"
+	"strconv"
 	"time"
 
 	"go.opentelemetry.io/collector/pdata/pcommon"
@@ -310,24 +311,24 @@ func accessMetricAggTemporality() pathGetSetter {
 			metric := ctx.(metricTransformContext).GetMetric()
 			switch metric.DataType() {
 			case pmetric.MetricDataTypeSum:
-				return metric.Sum().AggregationTemporality()
+				return int64(metric.Sum().AggregationTemporality())
 			case pmetric.MetricDataTypeHistogram:
-				return metric.Histogram().AggregationTemporality()
+				return int64(metric.Histogram().AggregationTemporality())
 			case pmetric.MetricDataTypeExponentialHistogram:
-				return metric.ExponentialHistogram().AggregationTemporality()
+				return int64(metric.ExponentialHistogram().AggregationTemporality())
 			}
 			return nil
 		},
 		setter: func(ctx common.TransformContext, val interface{}) {
-			if newAggTemporality, ok := val.(pmetric.MetricAggregationTemporality); ok {
+			if newAggTemporality, ok := val.(int64); ok {
 				metric := ctx.(metricTransformContext).GetMetric()
 				switch metric.DataType() {
 				case pmetric.MetricDataTypeSum:
-					metric.Sum().SetAggregationTemporality(newAggTemporality)
+					metric.Sum().SetAggregationTemporality(pmetric.MetricAggregationTemporality(newAggTemporality))
 				case pmetric.MetricDataTypeHistogram:
-					metric.Histogram().SetAggregationTemporality(newAggTemporality)
+					metric.Histogram().SetAggregationTemporality(pmetric.MetricAggregationTemporality(newAggTemporality))
 				case pmetric.MetricDataTypeExponentialHistogram:
-					metric.ExponentialHistogram().SetAggregationTemporality(newAggTemporality)
+					metric.ExponentialHistogram().SetAggregationTemporality(pmetric.MetricAggregationTemporality(newAggTemporality))
 				}
 			}
 		},
@@ -340,16 +341,18 @@ func accessMetricIsMonotonic() pathGetSetter {
 			metric := ctx.(metricTransformContext).GetMetric()
 			switch metric.DataType() {
 			case pmetric.MetricDataTypeSum:
-				return metric.Sum().IsMonotonic()
+				return strconv.FormatBool(metric.Sum().IsMonotonic())
 			}
 			return nil
 		},
 		setter: func(ctx common.TransformContext, val interface{}) {
-			if newIsMonotonic, ok := val.(bool); ok {
+			if newIsMonotonic, ok := val.(string); ok {
 				metric := ctx.(metricTransformContext).GetMetric()
 				switch metric.DataType() {
 				case pmetric.MetricDataTypeSum:
-					metric.Sum().SetIsMonotonic(newIsMonotonic)
+					if boolVal, err := strconv.ParseBool(newIsMonotonic); err == nil {
+						metric.Sum().SetIsMonotonic(boolVal)
+					}
 				}
 			}
 		},

processor/transformprocessor/internal/metrics/metrics_test.go

@@ -1532,8 +1532,8 @@ func Test_newPathGetSetter_Metric(t *testing.T) {
 					Name: "aggregation_temporality",
 				},
 			},
-			orig: pmetric.MetricAggregationTemporalityCumulative,
-			new:  pmetric.MetricAggregationTemporalityDelta,
+			orig: int64(2),
+			new:  int64(1),
 			modified: func(metric pmetric.Metric) {
 				metric.Sum().SetAggregationTemporality(pmetric.MetricAggregationTemporalityDelta)
 			},
@@ -1548,8 +1548,8 @@ func Test_newPathGetSetter_Metric(t *testing.T) {
 					Name: "is_monotonic",
 				},
 			},
-			orig: true,
-			new:  false,
+			orig: "true",
+			new:  "false",
 			modified: func(metric pmetric.Metric) {
 				metric.Sum().SetIsMonotonic(false)
 			},

processor/transformprocessor/internal/metrics/processor_test.go

@@ -80,6 +80,19 @@ func TestProcess(t *testing.T) {
 				td.ResourceMetrics().At(0).ScopeMetrics().At(0).Metrics().At(0).SetDescription("Sum")
 			},
 		},
+		{
+			query: `set(metric.aggregation_temporality, 1) where metric.aggregation_temporality == 0`,
+			want: func(td pmetric.Metrics) {
+				td.ResourceMetrics().At(0).ScopeMetrics().At(0).Metrics().At(0).Sum().SetAggregationTemporality(pmetric.MetricAggregationTemporalityDelta)
+				td.ResourceMetrics().At(0).ScopeMetrics().At(0).Metrics().At(1).Histogram().SetAggregationTemporality(pmetric.MetricAggregationTemporalityDelta)
+			},
+		},
+		{
+			query: `set(metric.is_monotonic, "true") where metric.is_monotonic == "false"`,
+			want: func(td pmetric.Metrics) {
+				td.ResourceMetrics().At(0).ScopeMetrics().At(0).Metrics().At(0).Sum().SetIsMonotonic(true)
+			},
+		},
 	}
 
 	for _, tt := range tests {