fix: Azure Pipelines Scaler uses correct endpoint when demands are set (

kedacore#4387) (kedacore#4401)

Co-authored-by: Jorge Turrado Ferrero <Jorge_turrado@hotmail.es>
Co-authored-by: Andy Ward <mortx@toothless.eldarrin.io>

CHANGELOG.md

@@ -54,6 +54,7 @@ To learn more about active deprecations, we recommend checking [GitHub Discussio
 ### Fixes
 
 - **AWS SQS Scaler**: Respect `scaleOnInFlight` value ([#4276](https://github.com/kedacore/keda/issue/4276))
+- **Azure Pipelines**: Fix for disallowing `$top` on query when using `meta.parentID` method ([#4397])
 
 ### Deprecations
 

pkg/scalers/azure_pipelines_scaler.go

@@ -332,7 +332,13 @@ func getAzurePipelineRequest(ctx context.Context, url string, metadata *azurePip
 }
 
 func (s *azurePipelinesScaler) GetAzurePipelinesQueueLength(ctx context.Context) (int64, error) {
-	url := fmt.Sprintf("%s/_apis/distributedtask/pools/%d/jobrequests?$top=%d", s.metadata.organizationURL, s.metadata.poolID, s.metadata.jobsToFetch)
+	// HotFix Issue (#4387), $top changes the format of the returned JSON
+	var url string
+	if s.metadata.parent != "" {
+		url = fmt.Sprintf("%s/_apis/distributedtask/pools/%d/jobrequests", s.metadata.organizationURL, s.metadata.poolID)
+	} else {
+		url = fmt.Sprintf("%s/_apis/distributedtask/pools/%d/jobrequests?$top=%d", s.metadata.organizationURL, s.metadata.poolID, s.metadata.jobsToFetch)
+	}
 	body, err := getAzurePipelineRequest(ctx, url, s.metadata, s.httpClient)
 	if err != nil {
 		return -1, err

tests/.env

@@ -9,9 +9,11 @@ TF_AZURE_APP_INSIGHTS_NAME=
 TF_AZURE_DATA_EXPLORER_DB=
 TF_AZURE_DATA_EXPLORER_ENDPOINT=
 AZURE_DEVOPS_BUILD_DEFINITION_ID=
+AZURE_DEVOPS_DEMAND_PARENT_BUILD_DEFINITION_ID=
 AZURE_DEVOPS_ORGANIZATION_URL=
 AZURE_DEVOPS_PAT=
 AZURE_DEVOPS_POOL_NAME=
+AZURE_DEVOPS_DEMAND_POOL_NAME=
 AZURE_DEVOPS_PROJECT=
 TF_AZURE_EVENTHBUS_MANAGEMENT_CONNECTION_STRING=
 TF_AZURE_KEYVAULT_URI=