Allow ingesting native histograms

CHANGELOG.md

@@ -3,6 +3,7 @@
 ## master / unreleased
 * [CHANGE] Upgrade Dockerfile Node version from 14x to 18x. #5906
 * [CHANGE] Ingester: Remove `-querier.query-store-for-labels-enabled` flag. Querying long-term store for labels is always enabled. #5984
+* [FEATURE] Ingester: Experimental: Enable native histogram ingestion via `-blocks-storage.tsdb.enable-native-histograms` flag. #5986
 * [ENHANCEMENT] rulers: Add support to persist tokens in rulers. #5987
 * [ENHANCEMENT] Query Frontend/Querier: Added store gateway postings touched count and touched size in Querier stats and log in Query Frontend. #5892
 * [ENHANCEMENT] Query Frontend/Querier: Returns `warnings` on prometheus query responses. #5916

docs/blocks-storage/querier.md

@@ -1441,4 +1441,8 @@ blocks_storage:
     # be out-of-order.
     # CLI flag: -blocks-storage.tsdb.out-of-order-cap-max
     [out_of_order_cap_max: <int> | default = 32]
+
+    # [EXPERIMENTAL] True to enable native histogram.
+    # CLI flag: -blocks-storage.tsdb.enable-native-histograms
+    [enable_native_histograms: <boolean> | default = false]
 ```

docs/blocks-storage/store-gateway.md

@@ -1566,4 +1566,8 @@ blocks_storage:
     # be out-of-order.
     # CLI flag: -blocks-storage.tsdb.out-of-order-cap-max
     [out_of_order_cap_max: <int> | default = 32]
+
+    # [EXPERIMENTAL] True to enable native histogram.
+    # CLI flag: -blocks-storage.tsdb.enable-native-histograms
+    [enable_native_histograms: <boolean> | default = false]
 ```

docs/configuration/config-file-reference.md

@@ -1998,6 +1998,10 @@ tsdb:
   # be out-of-order.
   # CLI flag: -blocks-storage.tsdb.out-of-order-cap-max
   [out_of_order_cap_max: <int> | default = 32]
+
+  # [EXPERIMENTAL] True to enable native histogram.
+  # CLI flag: -blocks-storage.tsdb.enable-native-histograms
+  [enable_native_histograms: <boolean> | default = false]
 ```
 
 ### `compactor_config`

docs/configuration/v1-guarantees.md

@@ -111,3 +111,5 @@ Currently experimental features are:
 - OTLP Receiver
 - Persistent tokens in the Ruler Ring:
   - `-ruler.ring.tokens-file-path` (path) CLI flag
+- Native Histograms
+  - Ingestion can be enabled by setting `-blocks-storage.tsdb.enable-native-histograms=true` on Ingester.

pkg/cortexpb/compat.go

@@ -18,7 +18,7 @@ import (
 	"github.com/cortexproject/cortex/pkg/util"
 )
 
-// ToWriteRequest converts matched slices of Labels, Samples and Metadata into a WriteRequest proto.
+// ToWriteRequest converts matched slices of Labels, Samples, Metadata and Histograms into a WriteRequest proto.
 // It gets timeseries from the pool, so ReuseSlice() should be called when done.
 func ToWriteRequest(lbls []labels.Labels, samples []Sample, metadata []*MetricMetadata, histograms []Histogram, source WriteRequest_SourceEnum) *WriteRequest {
 	req := &WriteRequest{
@@ -27,13 +27,17 @@ func ToWriteRequest(lbls []labels.Labels, samples []Sample, metadata []*MetricMe
 		Source:     source,
 	}
 
-	for i, s := range samples {
+	i := 0
+	for i < len(samples) || i < len(histograms) {
 		ts := TimeseriesFromPool()
 		ts.Labels = append(ts.Labels, FromLabelsToLabelAdapters(lbls[i])...)
-		ts.Samples = append(ts.Samples, s)
+		if i < len(samples) {
+			ts.Samples = append(ts.Samples, samples[i])
+		}
 		if i < len(histograms) {
 			ts.Histograms = append(ts.Histograms, histograms[i])
 		}
+		i++
 		req.Timeseries = append(req.Timeseries, PreallocTimeseries{TimeSeries: ts})
 	}
 

pkg/cortexpb/histograms.go

@@ -0,0 +1,120 @@
+// Copyright 2017 The Prometheus Authors
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package cortexpb
+
+import "github.com/prometheus/prometheus/model/histogram"
+
+func (h Histogram) IsFloatHistogram() bool {
+	_, ok := h.GetCount().(*Histogram_CountFloat)
+	return ok
+}
+
+// HistogramProtoToHistogram extracts a (normal integer) Histogram from the
+// provided proto message. The caller has to make sure that the proto message
+// represents an interger histogram and not a float histogram.
+// Changed from https://github.com/prometheus/prometheus/blob/0ab95536115adfe50af249d36d73674be694ca3f/storage/remote/codec.go#L626-L645
+func HistogramProtoToHistogram(hp Histogram) *histogram.Histogram {
+	if hp.IsFloatHistogram() {
+		panic("HistogramProtoToHistogram called with a float histogram")
+	}
+	return &histogram.Histogram{
+		CounterResetHint: histogram.CounterResetHint(hp.ResetHint),
+		Schema:           hp.Schema,
+		ZeroThreshold:    hp.ZeroThreshold,
+		ZeroCount:        hp.GetZeroCountInt(),
+		Count:            hp.GetCountInt(),
+		Sum:              hp.Sum,
+		PositiveSpans:    spansProtoToSpans(hp.GetPositiveSpans()),
+		PositiveBuckets:  hp.GetPositiveDeltas(),
+		NegativeSpans:    spansProtoToSpans(hp.GetNegativeSpans()),
+		NegativeBuckets:  hp.GetNegativeDeltas(),
+	}
+}
+
+// FloatHistogramProtoToFloatHistogram extracts a float Histogram from the provided proto message.
+// The caller has to make sure that the proto message represents a float histogram and not an
+// integer histogram, or it panics.
+// Changed from https://github.com/prometheus/prometheus/blob/0ab95536115adfe50af249d36d73674be694ca3f/storage/remote/codec.go#L647-L667
+func FloatHistogramProtoToFloatHistogram(hp Histogram) *histogram.FloatHistogram {
+	if !hp.IsFloatHistogram() {
+		panic("FloatHistogramProtoToFloatHistogram called with an integer histogram")
+	}
+	return &histogram.FloatHistogram{
+		CounterResetHint: histogram.CounterResetHint(hp.ResetHint),
+		Schema:           hp.Schema,
+		ZeroThreshold:    hp.ZeroThreshold,
+		ZeroCount:        hp.GetZeroCountFloat(),
+		Count:            hp.GetCountFloat(),
+		Sum:              hp.Sum,
+		PositiveSpans:    spansProtoToSpans(hp.GetPositiveSpans()),
+		PositiveBuckets:  hp.GetPositiveCounts(),
+		NegativeSpans:    spansProtoToSpans(hp.GetNegativeSpans()),
+		NegativeBuckets:  hp.GetNegativeCounts(),
+	}
+}
+
+// HistogramToHistogramProto converts a (normal integer) Histogram to its protobuf message type.
+// Changed from https://github.com/prometheus/prometheus/blob/0ab95536115adfe50af249d36d73674be694ca3f/storage/remote/codec.go#L709-L723
+func HistogramToHistogramProto(timestamp int64, h *histogram.Histogram) Histogram {
+	return Histogram{
+		Count:          &Histogram_CountInt{CountInt: h.Count},
+		Sum:            h.Sum,
+		Schema:         h.Schema,
+		ZeroThreshold:  h.ZeroThreshold,
+		ZeroCount:      &Histogram_ZeroCountInt{ZeroCountInt: h.ZeroCount},
+		NegativeSpans:  spansToSpansProto(h.NegativeSpans),
+		NegativeDeltas: h.NegativeBuckets,
+		PositiveSpans:  spansToSpansProto(h.PositiveSpans),
+		PositiveDeltas: h.PositiveBuckets,
+		ResetHint:      Histogram_ResetHint(h.CounterResetHint),
+		TimestampMs:    timestamp,
+	}
+}
+
+// FloatHistogramToHistogramProto converts a float Histogram to a normal
+// Histogram's protobuf message type.
+// Changed from https://github.com/prometheus/prometheus/blob/0ab95536115adfe50af249d36d73674be694ca3f/storage/remote/codec.go#L725-L739
+func FloatHistogramToHistogramProto(timestamp int64, fh *histogram.FloatHistogram) Histogram {
+	return Histogram{
+		Count:          &Histogram_CountFloat{CountFloat: fh.Count},
+		Sum:            fh.Sum,
+		Schema:         fh.Schema,
+		ZeroThreshold:  fh.ZeroThreshold,
+		ZeroCount:      &Histogram_ZeroCountFloat{ZeroCountFloat: fh.ZeroCount},
+		NegativeSpans:  spansToSpansProto(fh.NegativeSpans),
+		NegativeCounts: fh.NegativeBuckets,
+		PositiveSpans:  spansToSpansProto(fh.PositiveSpans),
+		PositiveCounts: fh.PositiveBuckets,
+		ResetHint:      Histogram_ResetHint(fh.CounterResetHint),
+		TimestampMs:    timestamp,
+	}
+}
+
+func spansProtoToSpans(s []BucketSpan) []histogram.Span {
+	spans := make([]histogram.Span, len(s))
+	for i := 0; i < len(s); i++ {
+		spans[i] = histogram.Span{Offset: s[i].Offset, Length: s[i].Length}
+	}
+
+	return spans
+}
+
+func spansToSpansProto(s []histogram.Span) []BucketSpan {
+	spans := make([]BucketSpan, len(s))
+	for i := 0; i < len(s); i++ {
+		spans[i] = BucketSpan{Offset: s[i].Offset, Length: s[i].Length}
+	}
+
+	return spans
+}

pkg/distributor/distributor.go

@@ -69,6 +69,9 @@ const (
 	// mergeSlicesParallelism is a constant of how much go routines we should use to merge slices, and
 	// it was based on empirical observation: See BenchmarkMergeSlicesParallel
 	mergeSlicesParallelism = 8
+
+	sampleMetricTypeFloat     = "float"
+	sampleMetricTypeHistogram = "histogram"
 )
 
 // Distributor is a storage.SampleAppender and a client.Querier which
@@ -276,7 +279,7 @@ func New(cfg Config, clientConfig ingester_client.Config, limits *validation.Ove
 			Namespace: "cortex",
 			Name:      "distributor_received_samples_total",
 			Help:      "The total number of received samples, excluding rejected and deduped samples.",
-		}, []string{"user"}),
+		}, []string{"user", "type"}),
 		receivedExemplars: promauto.With(reg).NewCounterVec(prometheus.CounterOpts{
 			Namespace: "cortex",
 			Name:      "distributor_received_exemplars_total",
@@ -291,7 +294,7 @@ func New(cfg Config, clientConfig ingester_client.Config, limits *validation.Ove
 			Namespace: "cortex",
 			Name:      "distributor_samples_in_total",
 			Help:      "The total number of samples that have come in to the distributor, including rejected or deduped samples.",
-		}, []string{"user"}),
+		}, []string{"user", "type"}),
 		incomingExemplars: promauto.With(reg).NewCounterVec(prometheus.CounterOpts{
 			Namespace: "cortex",
 			Name:      "distributor_exemplars_in_total",
@@ -428,10 +431,12 @@ func (d *Distributor) cleanupInactiveUser(userID string) {
 
 	d.HATracker.CleanupHATrackerMetricsForUser(userID)
 
-	d.receivedSamples.DeleteLabelValues(userID)
+	d.receivedSamples.DeleteLabelValues(userID, sampleMetricTypeFloat)
+	d.receivedSamples.DeleteLabelValues(userID, sampleMetricTypeHistogram)
 	d.receivedExemplars.DeleteLabelValues(userID)
 	d.receivedMetadata.DeleteLabelValues(userID)
-	d.incomingSamples.DeleteLabelValues(userID)
+	d.incomingSamples.DeleteLabelValues(userID, sampleMetricTypeFloat)
+	d.incomingSamples.DeleteLabelValues(userID, sampleMetricTypeHistogram)
 	d.incomingExemplars.DeleteLabelValues(userID)
 	d.incomingMetadata.DeleteLabelValues(userID)
 	d.nonHASamples.DeleteLabelValues(userID)
@@ -547,7 +552,7 @@ func (d *Distributor) validateSeries(ts cortexpb.PreallocTimeseries, userID stri
 		// Only alloc when data present
 		samples = make([]cortexpb.Sample, 0, len(ts.Samples))
 		for _, s := range ts.Samples {
-			if err := validation.ValidateSample(d.validateMetrics, limits, userID, ts.Labels, s); err != nil {
+			if err := validation.ValidateSampleTimestamp(d.validateMetrics, limits, userID, ts.Labels, s.TimestampMs); err != nil {
 				return emptyPreallocSeries, err
 			}
 			samples = append(samples, s)
@@ -574,8 +579,13 @@ func (d *Distributor) validateSeries(ts cortexpb.PreallocTimeseries, userID stri
 	if len(ts.Histograms) > 0 {
 		// Only alloc when data present
 		histograms = make([]cortexpb.Histogram, 0, len(ts.Histograms))
-		// TODO(yeya24): we need to have validations for native histograms
-		// at some point. Skip validations for now.
+		for _, h := range ts.Histograms {
+			// TODO(yeya24): add other validations for native histogram.
+			// For example, Prometheus scrape has bucket limit and schema check.
+			if err := validation.ValidateSampleTimestamp(d.validateMetrics, limits, userID, ts.Labels, h.TimestampMs); err != nil {
+				return emptyPreallocSeries, err
+			}
+		}
 		histograms = append(histograms, ts.Histograms...)
 	}
 
@@ -607,14 +617,17 @@ func (d *Distributor) Push(ctx context.Context, req *cortexpb.WriteRequest) (*co
 	now := time.Now()
 	d.activeUsers.UpdateUserTimestamp(userID, now)
 
-	numSamples := 0
+	numFloatSamples := 0
+	numHistogramSamples := 0
 	numExemplars := 0
 	for _, ts := range req.Timeseries {
-		numSamples += len(ts.Samples) + len(ts.Histograms)
+		numFloatSamples += len(ts.Samples)
+		numHistogramSamples += len(ts.Histograms)
 		numExemplars += len(ts.Exemplars)
 	}
 	// Count the total samples, exemplars in, prior to validation or deduplication, for comparison with other metrics.
-	d.incomingSamples.WithLabelValues(userID).Add(float64(numSamples))
+	d.incomingSamples.WithLabelValues(userID, sampleMetricTypeFloat).Add(float64(numFloatSamples))
+	d.incomingSamples.WithLabelValues(userID, sampleMetricTypeHistogram).Add(float64(numFloatSamples))
 	d.incomingExemplars.WithLabelValues(userID).Add(float64(numExemplars))
 	// Count the total number of metadata in.
 	d.incomingMetadata.WithLabelValues(userID).Add(float64(len(req.Metadata)))
@@ -642,31 +655,32 @@ func (d *Distributor) Push(ctx context.Context, req *cortexpb.WriteRequest) (*co
 
 			if errors.Is(err, ha.ReplicasNotMatchError{}) {
 				// These samples have been deduped.
-				d.dedupedSamples.WithLabelValues(userID, cluster).Add(float64(numSamples))
+				d.dedupedSamples.WithLabelValues(userID, cluster).Add(float64(numFloatSamples + numHistogramSamples))
 				return nil, httpgrpc.Errorf(http.StatusAccepted, err.Error())
 			}
 
 			if errors.Is(err, ha.TooManyReplicaGroupsError{}) {
-				d.validateMetrics.DiscardedSamples.WithLabelValues(validation.TooManyHAClusters, userID).Add(float64(numSamples))
+				d.validateMetrics.DiscardedSamples.WithLabelValues(validation.TooManyHAClusters, userID).Add(float64(numFloatSamples + numHistogramSamples))
 				return nil, httpgrpc.Errorf(http.StatusBadRequest, err.Error())
 			}
 
 			return nil, err
 		}
 		// If there wasn't an error but removeReplica is false that means we didn't find both HA labels.
 		if !removeReplica {
-			d.nonHASamples.WithLabelValues(userID).Add(float64(numSamples))
+			d.nonHASamples.WithLabelValues(userID).Add(float64(numFloatSamples + numHistogramSamples))
 		}
 	}
 
 	// A WriteRequest can only contain series or metadata but not both. This might change in the future.
-	seriesKeys, validatedTimeseries, validatedSamples, validatedExemplars, firstPartialErr, err := d.prepareSeriesKeys(ctx, req, userID, limits, removeReplica)
+	seriesKeys, validatedTimeseries, validatedFloatSamples, validatedHistogramSamples, validatedExemplars, firstPartialErr, err := d.prepareSeriesKeys(ctx, req, userID, limits, removeReplica)
 	if err != nil {
 		return nil, err
 	}
 	metadataKeys, validatedMetadata, firstPartialErr := d.prepareMetadataKeys(req, limits, userID, firstPartialErr)
 
-	d.receivedSamples.WithLabelValues(userID).Add(float64(validatedSamples))
+	d.receivedSamples.WithLabelValues(userID, sampleMetricTypeFloat).Add(float64(validatedFloatSamples))
+	d.receivedSamples.WithLabelValues(userID, sampleMetricTypeHistogram).Add(float64(validatedHistogramSamples))
 	d.receivedExemplars.WithLabelValues(userID).Add(float64(validatedExemplars))
 	d.receivedMetadata.WithLabelValues(userID).Add(float64(len(validatedMetadata)))
 
@@ -677,18 +691,19 @@ func (d *Distributor) Push(ctx context.Context, req *cortexpb.WriteRequest) (*co
 		return &cortexpb.WriteResponse{}, firstPartialErr
 	}
 
-	totalN := validatedSamples + validatedExemplars + len(validatedMetadata)
+	totalSamples := validatedFloatSamples + validatedHistogramSamples
+	totalN := totalSamples + validatedExemplars + len(validatedMetadata)
 	if !d.ingestionRateLimiter.AllowN(now, userID, totalN) {
 		// Ensure the request slice is reused if the request is rate limited.
 		cortexpb.ReuseSlice(req.Timeseries)
 
-		d.validateMetrics.DiscardedSamples.WithLabelValues(validation.RateLimited, userID).Add(float64(validatedSamples))
+		d.validateMetrics.DiscardedSamples.WithLabelValues(validation.RateLimited, userID).Add(float64(totalSamples))
 		d.validateMetrics.DiscardedExemplars.WithLabelValues(validation.RateLimited, userID).Add(float64(validatedExemplars))
 		d.validateMetrics.DiscardedMetadata.WithLabelValues(validation.RateLimited, userID).Add(float64(len(validatedMetadata)))
 		// Return a 429 here to tell the client it is going too fast.
 		// Client may discard the data or slow down and re-send.
 		// Prometheus v2.26 added a remote-write option 'retry_on_http_429'.
-		return nil, httpgrpc.Errorf(http.StatusTooManyRequests, "ingestion rate limit (%v) exceeded while adding %d samples and %d metadata", d.ingestionRateLimiter.Limit(now, userID), validatedSamples, len(validatedMetadata))
+		return nil, httpgrpc.Errorf(http.StatusTooManyRequests, "ingestion rate limit (%v) exceeded while adding %d samples and %d metadata", d.ingestionRateLimiter.Limit(now, userID), totalSamples, len(validatedMetadata))
 	}
 
 	// totalN included samples and metadata. Ingester follows this pattern when computing its ingestion rate.
@@ -810,15 +825,16 @@ func (d *Distributor) prepareMetadataKeys(req *cortexpb.WriteRequest, limits *va
 	return metadataKeys, validatedMetadata, firstPartialErr
 }
 
-func (d *Distributor) prepareSeriesKeys(ctx context.Context, req *cortexpb.WriteRequest, userID string, limits *validation.Limits, removeReplica bool) ([]uint32, []cortexpb.PreallocTimeseries, int, int, error, error) {
+func (d *Distributor) prepareSeriesKeys(ctx context.Context, req *cortexpb.WriteRequest, userID string, limits *validation.Limits, removeReplica bool) ([]uint32, []cortexpb.PreallocTimeseries, int, int, int, error, error) {
 	pSpan, _ := opentracing.StartSpanFromContext(ctx, "prepareSeriesKeys")
 	defer pSpan.Finish()
 
 	// For each timeseries or samples, we compute a hash to distribute across ingesters;
 	// check each sample/metadata and discard if outside limits.
 	validatedTimeseries := make([]cortexpb.PreallocTimeseries, 0, len(req.Timeseries))
 	seriesKeys := make([]uint32, 0, len(req.Timeseries))
-	validatedSamples := 0
+	validatedFloatSamples := 0
+	validatedHistogramSamples := 0
 	validatedExemplars := 0
 
 	var firstPartialErr error
@@ -839,7 +855,9 @@ func (d *Distributor) prepareSeriesKeys(ctx context.Context, req *cortexpb.Write
 		if len(ts.Samples) > 0 {
 			latestSampleTimestampMs = max(latestSampleTimestampMs, ts.Samples[len(ts.Samples)-1].TimestampMs)
 		}
-		// TODO(yeya24): use timestamp of the latest native histogram in the series as well.
+		if len(ts.Histograms) > 0 {
+			latestSampleTimestampMs = max(latestSampleTimestampMs, ts.Histograms[len(ts.Histograms)-1].TimestampMs)
+		}
 
 		if mrc := limits.MetricRelabelConfigs; len(mrc) > 0 {
 			l, _ := relabel.Process(cortexpb.FromLabelAdaptersToLabels(ts.Labels), mrc...)
@@ -885,7 +903,7 @@ func (d *Distributor) prepareSeriesKeys(ctx context.Context, req *cortexpb.Write
 		// label and dropped labels (if any)
 		key, err := d.tokenForLabels(userID, ts.Labels)
 		if err != nil {
-			return nil, nil, 0, 0, nil, err
+			return nil, nil, 0, 0, 0, nil, err
 		}
 		validatedSeries, validationErr := d.validateSeries(ts, userID, skipLabelNameValidation, limits)
 
@@ -904,11 +922,11 @@ func (d *Distributor) prepareSeriesKeys(ctx context.Context, req *cortexpb.Write
 
 		seriesKeys = append(seriesKeys, key)
 		validatedTimeseries = append(validatedTimeseries, validatedSeries)
-		// TODO(yeya24): add histogram samples as well when supported.
-		validatedSamples += len(ts.Samples)
+		validatedFloatSamples += len(ts.Samples)
+		validatedHistogramSamples += len(ts.Histograms)
 		validatedExemplars += len(ts.Exemplars)
 	}
-	return seriesKeys, validatedTimeseries, validatedSamples, validatedExemplars, firstPartialErr, nil
+	return seriesKeys, validatedTimeseries, validatedFloatSamples, validatedHistogramSamples, validatedExemplars, firstPartialErr, nil
 }
 
 func sortLabelsIfNeeded(labels []cortexpb.LabelAdapter) {