Prometheus: Remove prometheusRunQueriesInParallel feature toggle (#103824)

* remove prometheusRunQueriesInParallel feature toggle

* fix the unit test

docs/sources/setup-grafana/configure-grafana/feature-toggles/index.md

@@ -32,7 +32,6 @@ Most [generally available](https://grafana.com/docs/release-life-cycle/#general-
 | `logsContextDatasourceUi`              | Allow datasource to provide custom UI for context view                                                                              | Yes                |
 | `lokiQuerySplitting`                   | Split large interval queries into subqueries with smaller time intervals                                                            | Yes                |
 | `influxdbBackendMigration`             | Query InfluxDB InfluxQL without the proxy                                                                                           | Yes                |
-| `prometheusRunQueriesInParallel`       | Enables running Prometheus queries in parallel                                                                                      | Yes                |
 | `dataplaneFrontendFallback`            | Support dataplane contract field name change for transformations and field name matchers where the name is different                | Yes                |
 | `unifiedRequestLog`                    | Writes error logs to the request logger                                                                                             | Yes                |
 | `pluginsDetailsRightPanel`             | Enables right panel for the plugins details page                                                                                    | Yes                |

packages/grafana-data/src/types/featureToggles.gen.ts

@@ -132,11 +132,6 @@ export interface FeatureToggles {
   */
   influxdbRunQueriesInParallel?: boolean;
   /**
-  * Enables running Prometheus queries in parallel
-  * @default true
-  */
-  prometheusRunQueriesInParallel?: boolean;
-  /**
   * Changes logs responses from Loki to be compliant with the dataplane specification.
   */
   lokiLogsDataplane?: boolean;

pkg/promlib/healthcheck_test.go

@@ -124,8 +124,10 @@ func getPluginContext() backend.PluginContext {
 		User:                       nil,
 		AppInstanceSettings:        nil,
 		DataSourceInstanceSettings: getPromInstanceSettings(),
+		GrafanaConfig:              backend.NewGrafanaCfg(map[string]string{"concurrent_query_count": "10"}),
 	}
 }
+
 func getPromInstanceSettings() *backend.DataSourceInstanceSettings {
 	return &backend.DataSourceInstanceSettings{
 		ID:                      0,

pkg/promlib/querydata/framing_test.go

@@ -7,6 +7,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"net/url"
 	"os"
 	"path/filepath"
 	"testing"
@@ -20,7 +21,7 @@ import (
 	"github.com/grafana/grafana/pkg/promlib/models"
 )
 
-var update = true
+var update = false
 
 func TestRangeResponses(t *testing.T) {
 	tt := []struct {
@@ -145,10 +146,42 @@ func runQuery(response []byte, q *backend.QueryDataRequest) (*backend.QueryDataR
 	if err != nil {
 		return nil, err
 	}
+
+	// Create initial response
 	res := &http.Response{
 		StatusCode: 200,
 		Body:       io.NopCloser(bytes.NewReader(response)),
+		Request: &http.Request{
+			URL: &url.URL{
+				Path: "api/v1/query_range",
+			},
+		},
 	}
-	tCtx.httpProvider.setResponse(res, res)
+
-	return tCtx.queryData.Execute(context.Background(), q)
+	// Create a proper clone for the exemplar response with a different path
+	exemplarRes := &http.Response{
+		StatusCode: 200,
+		Body:       io.NopCloser(bytes.NewReader(response)),
+		Request: &http.Request{
+			URL: &url.URL{
+				Path: "api/v1/query_exemplars",
+			},
+		},
+	}
+
+	tCtx.httpProvider.setResponse(res, exemplarRes)
+
+	// Add GrafanaConfig to the context to prevent nil pointer dereference
+	ctx := backend.WithGrafanaConfig(context.Background(), backend.NewGrafanaCfg(map[string]string{
+		"concurrent_query_count": "10",
+	}))
+
+	// Add a PluginContext with GrafanaConfig to the request
+	q.PluginContext = backend.PluginContext{
+		GrafanaConfig: backend.NewGrafanaCfg(map[string]string{
+			"concurrent_query_count": "10",
+		}),
+	}
+
+	return tCtx.queryData.Execute(ctx, q)
 }

pkg/promlib/querydata/request.go

@@ -102,11 +102,6 @@ func (s *QueryData) Execute(ctx context.Context, req *backend.QueryDataRequest)
 
 	var (
 		hasPromQLScopeFeatureFlag = s.featureToggles.IsEnabled("promQLScope")
-		hasPrometheusRunQueriesInParallel = s.featureToggles.IsEnabled("prometheusRunQueriesInParallel")
-	)
-
-	if hasPrometheusRunQueriesInParallel {
-		var (
 		m                         sync.Mutex
 	)
 
@@ -118,7 +113,7 @@ func (s *QueryData) Execute(ctx context.Context, req *backend.QueryDataRequest)
 
 	_ = concurrency.ForEachJob(ctx, len(req.Queries), concurrentQueryCount, func(ctx context.Context, idx int) error {
 		query := req.Queries[idx]
-			r := s.handleQuery(ctx, query, fromAlert, hasPromQLScopeFeatureFlag, true)
+		r := s.handleQuery(ctx, query, fromAlert, hasPromQLScopeFeatureFlag)
 		if r != nil {
 			m.Lock()
 			result.Responses[query.RefID] = *r
@@ -126,20 +121,12 @@ func (s *QueryData) Execute(ctx context.Context, req *backend.QueryDataRequest)
 		}
 		return nil
 	})
-	} else {
-		for _, q := range req.Queries {
-			r := s.handleQuery(ctx, q, fromAlert, hasPromQLScopeFeatureFlag, false)
-			if r != nil {
-				result.Responses[q.RefID] = *r
-			}
-		}
-	}
 
 	return &result, nil
 }
 
 func (s *QueryData) handleQuery(ctx context.Context, bq backend.DataQuery, fromAlert,
-	hasPromQLScopeFeatureFlag, hasPrometheusRunQueriesInParallel bool) *backend.DataResponse {
+	hasPromQLScopeFeatureFlag bool) *backend.DataResponse {
 	traceCtx, span := s.tracer.Start(ctx, "datasource.prometheus")
 	defer span.End()
 	query, err := models.Parse(span, bq, s.TimeInterval, s.intervalCalculator, fromAlert, hasPromQLScopeFeatureFlag)
@@ -149,14 +136,14 @@ func (s *QueryData) handleQuery(ctx context.Context, bq backend.DataQuery, fromA
 		}
 	}
 
-	r := s.fetch(traceCtx, s.client, query, hasPrometheusRunQueriesInParallel)
+	r := s.fetch(traceCtx, s.client, query)
 	if r == nil {
 		s.log.FromContext(ctx).Debug("Received nil response from runQuery", "query", query.Expr)
 	}
 	return r
 }
 
-func (s *QueryData) fetch(traceCtx context.Context, client *client.Client, q *models.Query, hasPrometheusRunQueriesInParallel bool) *backend.DataResponse {
+func (s *QueryData) fetch(traceCtx context.Context, client *client.Client, q *models.Query) *backend.DataResponse {
 	logger := s.log.FromContext(traceCtx)
 	logger.Debug("Sending query", "start", q.Start, "end", q.End, "step", q.Step, "query", q.Expr /*, "queryTimeout", s.QueryTimeout*/)
 
@@ -171,7 +158,6 @@ func (s *QueryData) fetch(traceCtx context.Context, client *client.Client, q *mo
 	)
 
 	if q.InstantQuery {
-		if hasPrometheusRunQueriesInParallel {
 		wg.Add(1)
 		go func() {
 			defer wg.Done()
@@ -180,14 +166,9 @@ func (s *QueryData) fetch(traceCtx context.Context, client *client.Client, q *mo
 			addDataResponse(&res, dr)
 			m.Unlock()
 		}()
-		} else {
-			res := s.instantQuery(traceCtx, client, q)
-			addDataResponse(&res, dr)
-		}
 	}
 
 	if q.RangeQuery {
-		if hasPrometheusRunQueriesInParallel {
 		wg.Add(1)
 		go func() {
 			defer wg.Done()
@@ -196,14 +177,9 @@ func (s *QueryData) fetch(traceCtx context.Context, client *client.Client, q *mo
 			addDataResponse(&res, dr)
 			m.Unlock()
 		}()
-		} else {
-			res := s.rangeQuery(traceCtx, client, q)
-			addDataResponse(&res, dr)
-		}
 	}
 
 	if q.ExemplarQuery {
-		if hasPrometheusRunQueriesInParallel {
 		wg.Add(1)
 		go func() {
 			defer wg.Done()
@@ -217,15 +193,6 @@ func (s *QueryData) fetch(traceCtx context.Context, client *client.Client, q *mo
 			dr.Frames = append(dr.Frames, res.Frames...)
 			m.Unlock()
 		}()
-		} else {
-			res := s.exemplarQuery(traceCtx, client, q)
-			if res.Error != nil {
-				// If exemplar query returns error, we want to only log it and
-				// continue with other results processing
-				logger.Error("Exemplar query failed", "query", q.Expr, "err", res.Error)
-			}
-			dr.Frames = append(dr.Frames, res.Frames...)
-		}
 	}
 	wg.Wait()
 

pkg/promlib/querydata/request_test.go

@@ -96,15 +96,26 @@ func TestPrometheus_parseTimeSeriesResponse(t *testing.T) {
 
 		// Test fields
 		require.Len(t, res, 2)
-		require.Equal(t, res[0].Name, "exemplar")
+		// Find the exemplar frame
-		require.Equal(t, res[0].Fields[0].Name, "Time")
+		var exemplarFrame *data.Frame
-		require.Equal(t, res[0].Fields[1].Name, "Value")
+		var rangeFrame *data.Frame
-		require.Len(t, res[0].Fields, 6)
+		for _, frame := range res {
+			if frame.Name == "exemplar" {
+				exemplarFrame = frame
+			} else {
+				rangeFrame = frame
+			}
+		}
+		require.NotNil(t, exemplarFrame)
+		require.NotNil(t, rangeFrame)
+		require.Equal(t, "Time", exemplarFrame.Fields[0].Name)
+		require.Equal(t, "Value", exemplarFrame.Fields[1].Name)
+		require.Len(t, exemplarFrame.Fields, 6)
 
 		// Test correct values (sampled to 2)
-		require.Equal(t, res[0].Fields[1].Len(), 4)
+		require.Equal(t, 4, exemplarFrame.Fields[1].Len())
-		require.Equal(t, res[0].Fields[1].At(0), 0.009545445)
+		require.Equal(t, 0.009545445, exemplarFrame.Fields[1].At(0))
-		require.Equal(t, res[0].Fields[1].At(3), 0.003535405)
+		require.Equal(t, 0.003535405, exemplarFrame.Fields[1].At(3))
 	})
 
 	t.Run("matrix response should be parsed normally", func(t *testing.T) {
@@ -379,6 +390,11 @@ func executeWithHeaders(tctx *testContext, query backend.DataQuery, rqr any, eqr
 	req := backend.QueryDataRequest{
 		Queries: []backend.DataQuery{query},
 		Headers: headers,
+		PluginContext: backend.PluginContext{
+			GrafanaConfig: backend.NewGrafanaCfg(map[string]string{
+				"concurrent_query_count": "10",
+			}),
+		},
 	}
 
 	rangeRes, err := toAPIResponse(rqr)
@@ -399,7 +415,12 @@ func executeWithHeaders(tctx *testContext, query backend.DataQuery, rqr any, eqr
 	}()
 	tctx.httpProvider.setResponse(rangeRes, exemplarRes)
 
-	res, err := tctx.queryData.Execute(context.Background(), &req)
+	// Create context with GrafanaConfig
+	ctx := backend.WithGrafanaConfig(context.Background(), backend.NewGrafanaCfg(map[string]string{
+		"concurrent_query_count": "10",
+	}))
+
+	res, err := tctx.queryData.Execute(ctx, &req)
 	if err != nil {
 		return nil, err
 	}
@@ -505,7 +526,37 @@ func (p *fakeHttpClientProvider) GetTransport(opts ...httpclient.Options) (http.
 
 func (p *fakeHttpClientProvider) setResponse(rangeRes *http.Response, exemplarRes *http.Response) {
 	p.rangeRes = rangeRes
-	p.exemplarRes = exemplarRes
+
+	// Create a proper clone manually ensuring we have a fresh response
+	if exemplarRes != nil {
+		bodyBytes, _ := io.ReadAll(exemplarRes.Body)
+		err := exemplarRes.Body.Close() // Close the original
+		if err != nil {
+			fmt.Println(fmt.Errorf("exemplarRes body close error: %v", err))
+			return
+		}
+
+		// Create a new request if the original has one
+		var newRequest *http.Request
+		if exemplarRes.Request != nil {
+			newRequest = &http.Request{
+				Method: exemplarRes.Request.Method,
+				URL:    exemplarRes.Request.URL,
+				Header: exemplarRes.Request.Header.Clone(),
+			}
+		}
+
+		// Create a new response with the same data but new body
+		p.exemplarRes = &http.Response{
+			StatusCode: exemplarRes.StatusCode,
+			Body:       io.NopCloser(bytes.NewReader(bodyBytes)),
+			Request:    newRequest,
+			Header:     exemplarRes.Header.Clone(),
+		}
+
+		// Reset the original body with a new reader
+		exemplarRes.Body = io.NopCloser(bytes.NewReader(bodyBytes))
+	}
 }
 
 func (p *fakeHttpClientProvider) RoundTrip(req *http.Request) (*http.Response, error) {

pkg/services/featuremgmt/registry.go

@@ -206,14 +206,6 @@ var (
 			FrontendOnly: false,
 			Owner:        grafanaPartnerPluginsSquad,
 		},
-		{
-			Name:         "prometheusRunQueriesInParallel",
-			Description:  "Enables running Prometheus queries in parallel",
-			Stage:        FeatureStageGeneralAvailability,
-			FrontendOnly: false,
-			Owner:        grafanaOSSBigTent,
-			Expression:   "true", // enabled by default
-		},
 		{
 			Name:        "lokiLogsDataplane",
 			Description: "Changes logs responses from Loki to be compliant with the dataplane specification.",

pkg/services/featuremgmt/toggles_gen.csv

@@ -25,7 +25,6 @@ individualCookiePreferences,experimental,@grafana/grafana-backend-group,false,fa
 influxdbBackendMigration,GA,@grafana/partner-datasources,false,false,true
 influxqlStreamingParser,experimental,@grafana/partner-datasources,false,false,false
 influxdbRunQueriesInParallel,privatePreview,@grafana/partner-datasources,false,false,false
-prometheusRunQueriesInParallel,GA,@grafana/oss-big-tent,false,false,false
 lokiLogsDataplane,experimental,@grafana/observability-logs,false,false,false
 dataplaneFrontendFallback,GA,@grafana/observability-metrics,false,false,true
 disableSSEDataplane,experimental,@grafana/observability-metrics,false,false,false

pkg/services/featuremgmt/toggles_gen.go

@@ -111,10 +111,6 @@ const (
 	// Enables running InfluxDB Influxql queries in parallel
 	FlagInfluxdbRunQueriesInParallel = "influxdbRunQueriesInParallel"
 
-	// FlagPrometheusRunQueriesInParallel
-	// Enables running Prometheus queries in parallel
-	FlagPrometheusRunQueriesInParallel = "prometheusRunQueriesInParallel"
-
 	// FlagLokiLogsDataplane
 	// Changes logs responses from Loki to be compliant with the dataplane specification.
 	FlagLokiLogsDataplane = "lokiLogsDataplane"

pkg/services/featuremgmt/toggles_gen.json

@@ -2583,7 +2583,8 @@
       "metadata": {
         "name": "prometheusRunQueriesInParallel",
         "resourceVersion": "1743693517832",
-        "creationTimestamp": "2024-08-12T12:31:39Z"
+        "creationTimestamp": "2024-08-12T12:31:39Z",
+        "deletionTimestamp": "2025-04-10T18:16:18Z"
       },
       "spec": {
         "description": "Enables running Prometheus queries in parallel",