[exporter/prometheusremotewrite] Collector crashes while Scraping Histogram Metrics

[ramyasree0299]

Describe the bug
I'm running Opentelemetry collector contrib v0.56.0 Windows Amd64 binary with OTLP Receiver in Windows machine and have a client application in Spring boot v2.6.9 using Micrometer Registry OTLP v1.9.2 to publish metrics to collector binary in OTLP Format. But as soon as client application starts publishing the metrics, the collector Crashes.
Related to #7149

Steps to reproduce

1. Start otel collector binary with provided config below.
2. From https://start.spring.io/ start a new Spring boot project with v2.6.9 and add below micrometer Dependency

    <dependency>
      <groupId>io.micrometer</groupId>
      <artifactId>micrometer-registry-otlp</artifactId>
      <version>1.9.2</version>
    </dependency>

3. Create a OTLPMeterRegistry Bean to start exposing the metrics in OTLP format and start the application.(The collector publish endpoint will be default i.e; http://localhost:4318/v1/metrics)

    @Bean
    public OtlpMeterRegistry otlpMeterRegistry(){
        return new OtlpMeterRegistry();
    }

What did you expect to see?
Metrics to be remote written to the specified remote write endpoint.

What did you see instead?
Collector crashes with below exception

panic: runtime error: index out of range [-1]

goroutine 138 [running]:
go.opentelemetry.io/collector/pdata/internal.ImmutableUInt64Slice.At(...)
        go.opentelemetry.io/collector/pdata@v0.56.0/internal/generated_immutable_slice.go:125
github.com/open-telemetry/opentelemetry-collector-contrib/pkg/translator/prometheusremotewrite.addSingleHistogramDataPoint({0xc0002df6d8?}, {0x4591ed?}, {0xc000b3acc0?}, {{0x0?, 0x40e2c5?}, 0xc000285aa0?}, 0x50?)
        github.com/open-telemetry/opentelemetry-collector-contrib/pkg/translator/prometheusremotewrite@v0.56.0/helper.go:346 +0xf65
github.com/open-telemetry/opentelemetry-collector-contrib/pkg/translator/prometheusremotewrite.FromMetrics({0x9?}, {{0x0?, 0xb3ebe1c?}, 0xc000285aa0?})
        github.com/open-telemetry/opentelemetry-collector-contrib/pkg/translator/prometheusremotewrite@v0.56.0/metrics_to_prw.go:88 +0x74a
github.com/open-telemetry/opentelemetry-collector-contrib/exporter/prometheusremotewriteexporter.(*prwExporter).PushMetrics(0xc000147b90, {0x6179b98, 0xc000b41080}, {0x6179bd0?})
        github.com/open-telemetry/opentelemetry-collector-contrib/exporter/prometheusremotewriteexporter@v0.56.0/exporter.go:137 +0xd3
go.opentelemetry.io/collector/exporter/exporterhelper.(*metricsRequest).export(0x6179bd0?, {0x6179b98?, 0xc000b41080?})
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/metrics.go:65 +0x34
go.opentelemetry.io/collector/exporter/exporterhelper.(*timeoutSender).send(0xc000462750, {0x618ecf0, 0xc0001624e0})
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/common.go:225 +0x96
go.opentelemetry.io/collector/exporter/exporterhelper.(*retrySender).send(0xc000147c20, {0x618ecf0, 0xc0001624e0})
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/queued_retry.go:176 +0x556
go.opentelemetry.io/collector/exporter/exporterhelper.(*metricsSenderWithObservability).send(0xc001018450, {0x618ecf0, 0xc0001624e0})
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/metrics.go:132 +0x88
go.opentelemetry.io/collector/exporter/exporterhelper.(*queuedRetrySender).start.func1({0x51e46e0?, 0xc0001624e0})
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/queued_retry_inmemory.go:119 +0x5c
go.opentelemetry.io/collector/exporter/exporterhelper/internal.consumerFunc.consume(0xc0003a7200?, {0x51e46e0?, 0xc0001624e0?})
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/internal/bounded_memory_queue.go:82 +0x2c
go.opentelemetry.io/collector/exporter/exporterhelper/internal.(*boundedMemoryQueue).StartConsumers.func2()
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/internal/bounded_memory_queue.go:69 +0xcd
created by go.opentelemetry.io/collector/exporter/exporterhelper/internal.(*boundedMemoryQueue).StartConsumers
        go.opentelemetry.io/collector@v0.56.0/exporter/exporterhelper/internal/bounded_memory_queue.go:63 +0x98

What version did you use?
Version: 0.56.0

What config did you use?
Config:

receivers:
  otlp:
    protocols:
      grpc:
      http:
  zipkin:

exporters:
  prometheusremotewrite:
    endpoint: "https://remotewriteurl"
  jaeger:
    endpoint: "some.jaeger.endpoint"
  prometheus:
    endpoint: "127.0.0.1:9091"
  file:
    path: otel-logs.txt

processors:
  batch:

extensions:
  health_check:
  pprof:
  zpages:

service:
  extensions: [ pprof, zpages, health_check ]
  pipelines:
    metrics:
      receivers: [ otlp ]
      exporters: [ prometheus, file, prometheusremotewrite ]
    traces:
      receivers: [ otlp,zipkin ]
      exporters: [ jaeger, file ]
  telemetry:
    logs:
      level: "debug"

Environment
OS: Windows

[github-actions]

Pinging code owners: @Aneurysm9. See Adding Labels via Comments if you do not have permissions to add labels yourself.

[codeboten]

Pinging @dashpole @bertysentry as well as owners of the translator

[bertysentry]

This is the offending code in function addSingleHistogramDataPoint() in /pkg/translator/prometheusremotewrite/helper.go:

	// add le=+Inf bucket
	infBucket := &prompb.Sample{
		Timestamp: time,
	}
	if pt.Flags().HasFlag(pmetric.MetricDataPointFlagNoRecordedValue) {
		infBucket.Value = math.Float64frombits(value.StaleNaN)
	} else {
		cumulativeCount += pt.BucketCounts().At(pt.BucketCounts().Len() - 1) // panic: runtime error: index out of range [-1]
		infBucket.Value = float64(cumulativeCount)
	}

This part of the code is out of my scope of expertise, I'm sorry! @Aneurysm9 @dmitryax Any idea?

[dashpole]

Looks like it can panic on histograms without any buckets

[dashpole]

Fix: #12854