[dbnode] Report a histogram of loaded docs per query

src/dbnode/storage/index.go

@@ -1359,6 +1359,8 @@ func (i *nsIndex) queryWithSpan(
 		}
 	}
 
+	i.metrics.loadedDocsPerQuery.RecordValue(float64(results.TotalDocsCount()))
+
 	state.Lock()
 	// Take reference to vars to return while locked.
 	exhaustive := state.exhaustive
@@ -1844,6 +1846,8 @@ type nsIndexMetrics struct {
 	insertEndToEndLatency        tally.Timer
 	blocksEvictedMutableSegments tally.Counter
 	blockMetrics                 nsIndexBlocksMetrics
+
+	loadedDocsPerQuery tally.Histogram
 }
 
 func newNamespaceIndexMetrics(
@@ -1889,6 +1893,10 @@ func newNamespaceIndexMetrics(
 			"insert-end-to-end-latency", iopts.TimerOptions()),
 		blocksEvictedMutableSegments: scope.Counter("blocks-evicted-mutable-segments"),
 		blockMetrics:                 newNamespaceIndexBlocksMetrics(opts, blocksScope),
+		loadedDocsPerQuery: scope.Histogram(
+			"loaded-docs-per-query",
+			tally.MustMakeExponentialValueBuckets(10, 2, 16),
+		),
 	}
 }
 

src/dbnode/storage/index/aggregate_results.go

@@ -41,7 +41,8 @@ type aggregatedResults struct {
 	nsID          ident.ID
 	aggregateOpts AggregateResultsOptions
 
-	resultsMap *AggregateResultsMap
+	resultsMap     *AggregateResultsMap
+	totalDocsCount int
 
 	idPool    ident.Pool
 	bytesPool pool.CheckedBytesPool
@@ -94,6 +95,7 @@ func (r *aggregatedResults) Reset(
 
 	// reset all keys in the map next
 	r.resultsMap.Reset()
+	r.totalDocsCount = 0
 
 	// NB: could do keys+value in one step but I'm trying to avoid
 	// using an internal method of a code-gen'd type.
@@ -104,6 +106,7 @@ func (r *aggregatedResults) AddDocuments(batch []doc.Document) (int, error) {
 	r.Lock()
 	err := r.addDocumentsBatchWithLock(batch)
 	size := r.resultsMap.Len()
+	r.totalDocsCount += len(batch)
 	r.Unlock()
 	return size, err
 }
@@ -298,6 +301,13 @@ func (r *aggregatedResults) Size() int {
 	return l
 }
 
+func (r *aggregatedResults) TotalDocsCount() int {
+	r.RLock()
+	count := r.totalDocsCount
+	r.RUnlock()
+	return count
+}
+
 func (r *aggregatedResults) Finalize() {
 	r.Reset(nil, AggregateResultsOptions{})
 	if r.pool == nil {

src/dbnode/storage/index/aggregate_results_test.go

@@ -55,10 +55,16 @@ func TestAggResultsInsertInvalid(t *testing.T) {
 	require.Error(t, err)
 	require.Equal(t, 0, size)
 
+	require.Equal(t, 0, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
+
 	dInvalid = genDoc("", "foo")
 	size, err = res.AddDocuments([]doc.Document{dInvalid})
 	require.Error(t, err)
 	require.Equal(t, 0, size)
+
+	require.Equal(t, 0, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
 }
 
 func TestAggResultsInsertEmptyTermValue(t *testing.T) {
@@ -67,6 +73,21 @@ func TestAggResultsInsertEmptyTermValue(t *testing.T) {
 	size, err := res.AddDocuments([]doc.Document{dValidEmptyTerm})
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
+
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
+}
+
+func TestAggResultsInsertBatchOfTwo(t *testing.T) {
+	res := NewAggregateResults(nil, AggregateResultsOptions{}, testOpts)
+	d1 := genDoc("d1", "")
+	d2 := genDoc("d2", "")
+	size, err := res.AddDocuments([]doc.Document{d1, d2})
+	require.NoError(t, err)
+	require.Equal(t, 2, size)
+
+	require.Equal(t, 2, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
 }
 
 func TestAggResultsTermOnlyInsert(t *testing.T) {
@@ -78,15 +99,24 @@ func TestAggResultsTermOnlyInsert(t *testing.T) {
 	require.Error(t, err)
 	require.Equal(t, 0, size)
 
+	require.Equal(t, 0, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
+
 	dInvalid = genDoc("", "foo")
 	size, err = res.AddDocuments([]doc.Document{dInvalid})
 	require.Error(t, err)
 	require.Equal(t, 0, size)
 
+	require.Equal(t, 0, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
+
 	valid := genDoc("foo", "")
 	size, err = res.AddDocuments([]doc.Document{valid})
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
+
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 3, res.TotalDocsCount())
 }
 
 func testAggResultsInsertIdempotency(t *testing.T, res AggregateResults) {
@@ -95,9 +125,15 @@ func testAggResultsInsertIdempotency(t *testing.T, res AggregateResults) {
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
 
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
+
 	size, err = res.AddDocuments([]doc.Document{dValid})
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
+
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
 }
 
 func TestAggResultsInsertIdempotency(t *testing.T) {

src/dbnode/storage/index/results.go

@@ -40,7 +40,8 @@ type results struct {
 	nsID ident.ID
 	opts QueryResultsOptions
 
-	resultsMap *ResultsMap
+	resultsMap     *ResultsMap
+	totalDocsCount int
 
 	idPool    ident.Pool
 	bytesPool pool.CheckedBytesPool
@@ -88,6 +89,7 @@ func (r *results) Reset(nsID ident.ID, opts QueryResultsOptions) {
 
 	// Reset all keys in the map next, this will finalize the keys.
 	r.resultsMap.Reset()
+	r.totalDocsCount = 0
 
 	// NB: could do keys+value in one step but I'm trying to avoid
 	// using an internal method of a code-gen'd type.
@@ -101,6 +103,7 @@ func (r *results) AddDocuments(batch []doc.Document) (int, error) {
 	r.Lock()
 	err := r.addDocumentsBatchWithLock(batch)
 	size := r.resultsMap.Len()
+	r.totalDocsCount += len(batch)
 	r.Unlock()
 	return size, err
 }
@@ -173,6 +176,13 @@ func (r *results) Size() int {
 	return v
 }
 
+func (r *results) TotalDocsCount() int {
+	r.RLock()
+	count := r.totalDocsCount
+	r.RUnlock()
+	return count
+}
+
 func (r *results) Finalize() {
 	// Reset locks so cannot hold onto lock for call to Finalize.
 	r.Reset(nil, QueryResultsOptions{})

src/dbnode/storage/index/results_test.go

@@ -59,6 +59,9 @@ func TestResultsInsertInvalid(t *testing.T) {
 	size, err := res.AddDocuments([]doc.Document{dInvalid})
 	require.Error(t, err)
 	require.Equal(t, 0, size)
+
+	require.Equal(t, 0, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
 }
 
 func TestResultsInsertIdempotency(t *testing.T) {
@@ -68,9 +71,27 @@ func TestResultsInsertIdempotency(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
 
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
+
 	size, err = res.AddDocuments([]doc.Document{dValid})
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
+
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
+}
+
+func TestResultsInsertBatchOfTwo(t *testing.T) {
+	res := NewQueryResults(nil, QueryResultsOptions{}, testOpts)
+	d1 := doc.Document{ID: []byte("d1")}
+	d2 := doc.Document{ID: []byte("d2")}
+	size, err := res.AddDocuments([]doc.Document{d1, d2})
+	require.NoError(t, err)
+	require.Equal(t, 2, size)
+
+	require.Equal(t, 2, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
 }
 
 func TestResultsFirstInsertWins(t *testing.T) {
@@ -80,6 +101,9 @@ func TestResultsFirstInsertWins(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
 
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 1, res.TotalDocsCount())
+
 	tags, ok := res.Map().Get(ident.StringID("abc"))
 	require.True(t, ok)
 	require.Equal(t, 0, tags.Remaining())
@@ -92,6 +116,9 @@ func TestResultsFirstInsertWins(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, 1, size)
 
+	require.Equal(t, 1, res.Size())
+	require.Equal(t, 2, res.TotalDocsCount())
+
 	tags, ok = res.Map().Get(ident.StringID("abc"))
 	require.True(t, ok)
 	require.Equal(t, 0, tags.Remaining())
@@ -171,6 +198,7 @@ func TestResultsReset(t *testing.T) {
 	require.False(t, ok)
 	require.Equal(t, 0, tags.Remaining())
 	require.Equal(t, 0, res.Size())
+	require.Equal(t, 0, res.TotalDocsCount())
 }
 
 func TestResultsResetNamespaceClones(t *testing.T) {
@@ -206,6 +234,7 @@ func TestFinalize(t *testing.T) {
 	tags, ok = res.Map().Get(ident.StringID("abc"))
 	require.False(t, ok)
 	require.Equal(t, 0, res.Size())
+	require.Equal(t, 0, res.TotalDocsCount())
 
 	for _, entry := range res.Map().Iter() {
 		id, _ := entry.Key(), entry.Value()

src/dbnode/storage/index/types.go

@@ -126,6 +126,9 @@ type BaseResults interface {
 	// Size returns the number of IDs tracked.
 	Size() int
 
+	// TotalDocsCount returns the total number of documents observed.
+	TotalDocsCount() int
+
 	// AddDocuments adds the batch of documents to the results set, it will
 	// take a copy of the bytes backing the documents so the original can be
 	// modified after this function returns without affecting the results map.