Describe your environment Describe any aspect of your environment relevant to the problem, including your Python version,
python 3.7.10
opentelemetry_api-1.14.0
opentelemetry_sdk-1.14.0
opentelemetry_exporter_prometheus-1.12.0rc1

Steps to reproduce
Create Histogram and expose it with prometheus-exporter

What is the expected behavior?
Histograms should consistently appear on /metrics endpoint

What is the actual behavior?
Histograms disappears from /metrics endpoint after request/scrape. A change to metrics (in my case would be query to application itself) will make histogram appear again until next request/scrape. This is not happening neither in prometheus_client nor in javascript opentelemetry prometheus exporter.

Additional context
This behaviour is not optimal when multiple prometheus servers are scrapping metrics.
I have both auto instrumented and manually instrumented histograms both of them are disappearing.

Edit:
Don't know if its actually related but I hope it may be helpful.
I have additional observation. When pushing Histograms to OTEL Collector I sporadically encounter following error:

prometheusexporter@v0.67.0/accumulator.go:105	failed to translate metric	{"kind": "exporter", "data_type": "metrics", "name": "prometheus", "data_type": "\u0000", "metric_name": "http.server.duration"}

But its not the case with JS implementation its just works without any errors.