[otel-targetallocator] Unable to pass scrape_protocols

[halleystar]

Component(s)

target allocator

What happened?

Description

In k8s, I use opentelemetry-operator and servicemonitor to collect metrics. I want to use opentelemetry's OTEL exponential histograms. But there is a bug in it that makes my purpose impossible to achieve.

Steps to Reproduce

My config is like this:

Expected Result

export OTEL exponential histograms.

Actual Result

export classic histograms.

Kubernetes Version

1.28.0

Operator version

v0.102.0

Collector version

v0.102.0

Environment information

Environment

OS: (e.g., "Ubuntu 20.04")
Compiler(if manually compiled): (e.g., "go 14.2")

Log output

No response

Additional context

No response

[swiatekm]

Could you go into more detail as to what bug you're experiencing, and why you believe you need to set scrape_protocols in the receiver configuration? They can be set on the ServiceMonitor, and this should work correctly. Is the issue that they can't be set globally?

[halleystar]

I want to add:

scrapeProtocols:
  - PrometheusProto
  - OpenMetricsText1.0.0
  - OpenMetricsText0.0.1
  - PrometheusText0.0.4

to use opentelemetry's OTEL exponential。Regardless of whether it is placed in servicemonitor in opentelemetry operator. when i put it in servicemonitor, it also invalid:

[swiatekm]

Thanks for the demonstration. I investigated, and the reason this doesn't work is that our version of prometheus-operator doesn't support setting this field yet. Support was added in prometheus-operator 0.72.0, and we're on 0.71.2.

The reason we can't upgrade yet is that this requires Go 1.22, which we can't require until Go 1.21 becomes unsupported, as per #2757. @jaronoff97 @pavolloffay I believe this is the first time this policy has blocked a user requested feature.

[sfc-gh-rscott]

I'm experiencing the same issue. It would be nice to have this fixed.