I am collecting OpenTelemetry (OTEL) metrics from my services, forwarding them to an OTEL Collector, and then sending them to Prometheus through the OTEL endpoint. Here is a sample OTEL Collector configuration I am using:

# Receivers
  receivers:
    otlp:
      protocols:
        http:
          endpoint: 0.0.0.0:4318
          auth:
            authenticator: bearertokenauth

  # Processors
  processors:
    batch: {}

  # Exporters
  exporters:
    otlphttp/logs:
      endpoint: "http://loki-write:3100/otlp"
      tls:
        insecure: true
    otlphttp/metrics:
      endpoint: "http://prometheus-server/api/v1/otlp"
      tls:
        insecure: true
    otlphttp/traces:
      endpoint: "http://tempo-distributor:4318" 
      tls:
        insecure: true

    # Pipelines
  service:
    extensions:
      - health_check
      - bearertokenauth
    pipelines:
      logs:
        receivers: [otlp]
        processors: [batch]
        exporters: [otlphttp/logs]
      metrics:
        receivers: [otlp]
        processors: [batch]
        exporters: [otlphttp/metrics]
      traces:
        receivers: [otlp]
        processors: [batch]
        exporters: [otlphttp/traces]

Problem:

I can view most metrics in Prometheus (via Grafana), but histogram metrics are missing. For example, metrics like http_server_request_duration_sum do not appear in Prometheus.

Troubleshooting Steps:

1. To rule out issues with the OTEL pipeline, I tested sending the same metrics to a VictoriaMetrics backend. In VictoriaMetrics, I can see all metrics, including histograms (e.g., http_server_request_duration_sum). The only difference is that . in metric names is not replaced with _.
2. I enabled native histograms in Prometheus by modifying the deployment configuration, but this did not resolve the issue.
3. I am using the latest version of Prometheus deployed via a Helm chart in an EKS cluster.

Additional Context:

• OTEL metrics, except histograms, are visible in Prometheus.
• The environment is Kubernetes-based, and Prometheus is integrated using a Helm chart.
• I have reviewed multiple articles and GitHub issues but have not found a solution to this problem.

Question:

What could be the reason for histogram metrics not appearing in Prometheus? Are there any specific configurations in the OTEL Collector, Prometheus, or Grafana that I need to check?