prometheus/util/convertnhcb/convertnhcb.go
György Krajcsovits eafe72a0d0
Some checks failed
CI / Go tests (push) Has been cancelled
CI / More Go tests (push) Has been cancelled
CI / Go tests with previous Go version (push) Has been cancelled
CI / UI tests (push) Has been cancelled
CI / Go tests on Windows (push) Has been cancelled
CI / Mixins tests (push) Has been cancelled
CI / Build Prometheus for common architectures (0) (push) Has been cancelled
CI / Build Prometheus for common architectures (1) (push) Has been cancelled
CI / Build Prometheus for common architectures (2) (push) Has been cancelled
CI / Build Prometheus for all architectures (0) (push) Has been cancelled
CI / Build Prometheus for all architectures (1) (push) Has been cancelled
CI / Build Prometheus for all architectures (10) (push) Has been cancelled
CI / Build Prometheus for all architectures (11) (push) Has been cancelled
CI / Build Prometheus for all architectures (2) (push) Has been cancelled
CI / Build Prometheus for all architectures (3) (push) Has been cancelled
CI / Build Prometheus for all architectures (4) (push) Has been cancelled
CI / Build Prometheus for all architectures (5) (push) Has been cancelled
CI / Build Prometheus for all architectures (6) (push) Has been cancelled
CI / Build Prometheus for all architectures (7) (push) Has been cancelled
CI / Build Prometheus for all architectures (8) (push) Has been cancelled
CI / Build Prometheus for all architectures (9) (push) Has been cancelled
CI / Check generated parser (push) Has been cancelled
CI / golangci-lint (push) Has been cancelled
CI / fuzzing (push) Has been cancelled
CI / codeql (push) Has been cancelled
CI / Report status of build Prometheus for all architectures (push) Has been cancelled
CI / Publish main branch artifacts (push) Has been cancelled
CI / Publish release artefacts (push) Has been cancelled
CI / Publish UI on npm Registry (push) Has been cancelled
perf(nhcb): optimize away most allocations in convertnhcb
In general aim for the happy case when the exposer lists the buckets
in ascending order.

Use Compact(2) to compact the result of nhcb convert.

This is more in line with how client_golang optimizes spans vs
buckets.
aef8aedb4b/prometheus/histogram.go (L1485)

Signed-off-by: György Krajcsovits <gyorgy.krajcsovits@grafana.com>
2024-10-28 08:34:54 +01:00

258 lines
7.3 KiB
Go

// Copyright 2024 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 convertnhcb
import (
"errors"
"fmt"
"math"
"sort"
"strings"
"github.com/prometheus/prometheus/model/histogram"
"github.com/prometheus/prometheus/model/labels"
)
var (
errNegativeBucketCount = errors.New("bucket count must be non-negative")
errNegativeCount = errors.New("count must be non-negative")
errCountMismatch = errors.New("count mismatch")
errCountNotCumulative = errors.New("count is not cumulative")
)
type tempHistogramBucket struct {
le float64
count float64
}
// TempHistogram is used to collect information about classic histogram
// samples incrementally before creating a histogram.Histogram or
// histogram.FloatHistogram based on the values collected.
type TempHistogram struct {
buckets []tempHistogramBucket
count float64
sum float64
err error
hasCount bool
}
// NewTempHistogram creates a new TempHistogram to
// collect information about classic histogram samples.
func NewTempHistogram() TempHistogram {
return TempHistogram{
buckets: make([]tempHistogramBucket, 0, 10),
}
}
func (h TempHistogram) Err() error {
return h.err
}
func (h *TempHistogram) Reset() {
h.buckets = h.buckets[:0]
h.count = 0
h.sum = 0
h.err = nil
h.hasCount = false
}
func (h *TempHistogram) SetBucketCount(boundary, count float64) error {
if h.err != nil {
return h.err
}
if count < 0 {
h.err = fmt.Errorf("%w: le=%g, count=%g", errNegativeBucketCount, boundary, count)
return h.err
}
// Assume that the elements are added in order.
switch {
case len(h.buckets) == 0:
h.buckets = append(h.buckets, tempHistogramBucket{le: boundary, count: count})
case h.buckets[len(h.buckets)-1].le < boundary:
// Happy case is "<".
if count < h.buckets[len(h.buckets)-1].count {
h.err = fmt.Errorf("%w: %g < %g", errCountNotCumulative, count, h.buckets[len(h.buckets)-1].count)
return h.err
}
h.buckets = append(h.buckets, tempHistogramBucket{le: boundary, count: count})
case h.buckets[len(h.buckets)-1].le == boundary:
// Ignore this, as it is a duplicate sample.
default:
// Find the correct position to insert.
i := sort.Search(len(h.buckets), func(i int) bool {
return h.buckets[i].le >= boundary
})
if h.buckets[i].le == boundary {
// Ignore this, as it is a duplicate sample.
return nil
}
if i > 0 && count < h.buckets[i-1].count {
h.err = fmt.Errorf("%w: %g < %g", errCountNotCumulative, count, h.buckets[i-1].count)
return h.err
}
if count > h.buckets[i].count {
h.err = fmt.Errorf("%w: %g > %g", errCountNotCumulative, count, h.buckets[i].count)
return h.err
}
// Insert at the correct position unless duplicate.
h.buckets = append(h.buckets, tempHistogramBucket{})
copy(h.buckets[i+1:], h.buckets[i:])
h.buckets[i] = tempHistogramBucket{le: boundary, count: count}
}
return nil
}
func (h *TempHistogram) SetCount(count float64) error {
if h.err != nil {
return h.err
}
if count < 0 {
h.err = fmt.Errorf("%w: count=%g", errNegativeCount, count)
return h.err
}
h.count = count
h.hasCount = true
return nil
}
func (h *TempHistogram) SetSum(sum float64) error {
if h.err != nil {
return h.err
}
h.sum = sum
return nil
}
func (h TempHistogram) Convert() (*histogram.Histogram, *histogram.FloatHistogram, error) {
if h.err != nil {
return nil, nil, h.err
}
if len(h.buckets) == 0 || h.buckets[len(h.buckets)-1].le != math.Inf(1) {
// No +Inf bucket.
if !h.hasCount && len(h.buckets) > 0 {
// No count either, so set count to the last known bucket's count.
h.count = h.buckets[len(h.buckets)-1].count
}
// Let the last bucket be +Inf with the overall count.
h.buckets = append(h.buckets, tempHistogramBucket{le: math.Inf(1), count: h.count})
}
if !h.hasCount {
h.count = h.buckets[len(h.buckets)-1].count
h.hasCount = true
}
for _, b := range h.buckets {
intCount := int64(math.Round(b.count))
if b.count != float64(intCount) {
return h.convertToFloatHistogram()
}
}
intCount := uint64(math.Round(h.count))
if h.count != float64(intCount) {
return h.convertToFloatHistogram()
}
return h.convertToIntegerHistogram(intCount)
}
func (h TempHistogram) convertToIntegerHistogram(count uint64) (*histogram.Histogram, *histogram.FloatHistogram, error) {
rh := &histogram.Histogram{
Schema: histogram.CustomBucketsSchema,
Count: count,
Sum: h.sum,
PositiveSpans: []histogram.Span{{Length: uint32(len(h.buckets))}},
PositiveBuckets: make([]int64, len(h.buckets)),
}
if len(h.buckets) > 1 {
rh.CustomValues = make([]float64, len(h.buckets)-1) // Not storing the last +Inf bucket.
}
prevCount := int64(0)
prevDelta := int64(0)
for i, b := range h.buckets {
// delta is the actual bucket count as the input is cumulative.
delta := int64(b.count) - prevCount
rh.PositiveBuckets[i] = delta - prevDelta
prevCount = int64(b.count)
prevDelta = delta
if b.le != math.Inf(1) {
rh.CustomValues[i] = b.le
}
}
if count != uint64(h.buckets[len(h.buckets)-1].count) {
h.err = fmt.Errorf("%w: count=%d != le=%g count=%g", errCountMismatch, count, h.buckets[len(h.buckets)-1].le, h.buckets[len(h.buckets)-1].count)
return nil, nil, h.err
}
return rh.Compact(2), nil, nil
}
func (h TempHistogram) convertToFloatHistogram() (*histogram.Histogram, *histogram.FloatHistogram, error) {
rh := &histogram.FloatHistogram{
Schema: histogram.CustomBucketsSchema,
Count: h.count,
Sum: h.sum,
PositiveSpans: []histogram.Span{{Length: uint32(len(h.buckets))}},
PositiveBuckets: make([]float64, len(h.buckets)),
}
if len(h.buckets) > 1 {
rh.CustomValues = make([]float64, len(h.buckets)-1) // Not storing the last +Inf bucket.
}
prevCount := 0.0
for i, b := range h.buckets {
rh.PositiveBuckets[i] = b.count - prevCount
prevCount = b.count
if b.le != math.Inf(1) {
rh.CustomValues[i] = b.le
}
}
if h.count != h.buckets[len(h.buckets)-1].count {
h.err = fmt.Errorf("%w: count=%g != le=%g count=%g", errCountMismatch, h.count, h.buckets[len(h.buckets)-1].le, h.buckets[len(h.buckets)-1].count)
return nil, nil, h.err
}
return nil, rh.Compact(0), nil
}
func GetHistogramMetricBase(m labels.Labels, suffix string) labels.Labels {
mName := m.Get(labels.MetricName)
return labels.NewBuilder(m).
Set(labels.MetricName, strings.TrimSuffix(mName, suffix)).
Del(labels.BucketLabel).
Labels()
}
// GetHistogramMetricBaseName removes the suffixes _bucket, _sum, _count from
// the metric name. We specifically do not remove the _created suffix as that
// should be removed by the caller.
func GetHistogramMetricBaseName(s string) string {
if r, ok := strings.CutSuffix(s, "_bucket"); ok {
return r
}
if r, ok := strings.CutSuffix(s, "_sum"); ok {
return r
}
if r, ok := strings.CutSuffix(s, "_count"); ok {
return r
}
return s
}