lib/storage: parallelize tag filters execution a bit

This should reduce execution time when a query contains multiple tag filters and each such filter matches big number of time series.

lib/storage/index_db.go

@@ -2807,7 +2807,7 @@ func (is *indexSearch) getMetricIDsForDateAndFilters(date uint64, tfs *TagFilter
 		return a.tf.Less(b.tf)
 	})
 
-	// Populate metricIDs with the first non-negative filter.
+	// Populate metricIDs for the first non-negative filter.
 	var tfsPostponed []*tagFilter
 	var metricIDs *uint64set.Set
 	maxDateMetrics := maxMetrics * 50
@@ -2859,36 +2859,67 @@ func (is *indexSearch) getMetricIDsForDateAndFilters(date uint64, tfs *TagFilter
 	}
 
 	// Intersect metricIDs with the rest of filters.
+	var mu sync.Mutex
+	var wg sync.WaitGroup
+	var errGlobal error
+	// Limit the number of concurrent goroutines for metricIDs filtering in the hope they reduce the number
+	// of matching metrics to quite low value, so the remaining filters could be matched by metricName.
+	concurrencyCh := make(chan struct{}, 2)
 	for i := range tfsRemainingWithCount {
 		tfWithCount := tfsRemainingWithCount[i]
-		if n := uint64(metricIDs.Len()); n < 1000 || (n < tfWithCount.count/maxIndexScanLoopsPerMetric && n < uint64(maxMetrics)/10) {
-			// It should be faster performing metricName match on the remaining filters
-			// instead of scanning big number of entries in the inverted index for these filters.
-			for i < len(tfsRemainingWithCount) {
-				tfsPostponed = append(tfsPostponed, tfsRemainingWithCount[i].tf)
-				i++
-			}
-			break
-		}
 		tf := tfWithCount.tf
-		m, err := is.getMetricIDsForDateTagFilter(tf, date, tfs.commonPrefix, maxDateMetrics)
+		wg.Add(1)
-		if err != nil {
+		go func() {
-			return nil, err
+			concurrencyCh <- struct{}{}
-		}
+			defer func() {
-		if m.Len() >= maxDateMetrics {
+				<-concurrencyCh
-			// Too many time series found by a single tag filter. Postpone applying this filter via metricName match.
+				wg.Done()
-			tfsPostponed = append(tfsPostponed, tf)
+			}()
-			continue
+			mu.Lock()
-		}
+			metricIDsLen := metricIDs.Len()
-		if tf.isNegative {
+			mu.Unlock()
-			metricIDs.Subtract(m)
+			if metricIDsLen == 0 {
-		} else {
+				// Short circuit - there is no need in applying the remaining filters to empty set.
-			metricIDs.Intersect(m)
+				return
-		}
+			}
-		if metricIDs.Len() == 0 {
+			if n := uint64(metricIDsLen); n < 1000 || (n < tfWithCount.count/maxIndexScanLoopsPerMetric && n < uint64(maxMetrics)/10) {
-			// Short circuit - there is no need in applying the remaining filters to empty set.
+				// It should be faster performing metricName match on the remaining filters
-			return nil, nil
+				// instead of scanning big number of entries in the inverted index for these filters.
-		}
+				mu.Lock()
+				tfsPostponed = append(tfsPostponed, tf)
+				mu.Unlock()
+				return
+			}
+			isLocal := is.db.getIndexSearch(is.deadline)
+			m, err := isLocal.getMetricIDsForDateTagFilter(tf, date, tfs.commonPrefix, maxDateMetrics)
+			isLocal.db.putIndexSearch(isLocal)
+			if err != nil {
+				mu.Lock()
+				if errGlobal == nil {
+					errGlobal = err
+				}
+				mu.Unlock()
+				return
+			}
+			if m.Len() >= maxDateMetrics {
+				// Too many time series found by a single tag filter. Postpone applying this filter via metricName match.
+				mu.Lock()
+				tfsPostponed = append(tfsPostponed, tf)
+				mu.Unlock()
+				return
+			}
+			mu.Lock()
+			if tf.isNegative {
+				metricIDs.Subtract(m)
+			} else {
+				metricIDs.Intersect(m)
+			}
+			mu.Unlock()
+		}()
+	}
+	wg.Wait()
+	if errGlobal != nil {
+		return nil, errGlobal
 	}
 	if len(tfsPostponed) > 0 {
 		if n := metricIDs.Len(); n > 50000 && n > maxMetrics/10 {