Logging sampler configuration not working properly (?)

[je-al]

Describe the bug
Logging sampling configuration seems to not be applied

What did you expect to see?
A constant logging rate

What did you see instead?
Logging sampling isn't working as intended, we get approximatelly 1k loglines per second

What version did you use?
otel/opentelemetry-collector-contrib:0.3.0

What config did you use?

exporters:
  logging:
    loglevel: info
    sampling_initial: 5
    sampling_thereafter: 1000
...

service:
  extensions: [health_check]
  pipelines:
    traces:
      receivers: [sapm]
      processors: [memory_limiter, batch, attributes/newenvironment, queued_retry]
      exporters: [logging, sapm]

Additional context
I'm not sure whether this is a problem with our configuration, but dropped traces from the queued_retry processor are flooding our logs.

[je-al]

I've since been disabused of the notion that this configuration actually affected logging for other parts of the pipeline...
So I'm guessing the ask (?) would be for a way to actually be able to sample the logs coming off of the queued_retry processor, because right now the only way to avoid the flood of logs is to have rules dropping them in some sort of logs collector, or actually lower (raise?) the loglevel globally to ERROR (since dropped messages log at WARN), neither of which seem that convenient.

[jan25]

Are you seeing raw trace data logged? or just one line per trace printing number of spans inside the trace?

[bogdandrutu]

No longer applies, I think it was fixed