Change processedSpans&processedLogs labels

[JoeCqupt]

when i use BatchLogRecordProcessor/BatchSpanProcessor.
i want to know dropped ratio and exported success ratio and exported fail ratio and etc. but exist metrics can not compute those.

i want achieve this goal by updating processedSpans&processedLogs labels .

dropped ratio: dropped/processed
exported success ratio: exported/(processed - dropped)
exported fail ratio: 1 - exported/(processed - dropped)

[codecov]

Codecov Report

Patch coverage: 100.00% and project coverage change: -0.01 ⚠️

Comparison is base (b0c5c04) 91.32% compared to head (2749de9) 91.31%.

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #5631      +/-   ##
============================================
- Coverage     91.32%   91.31%   -0.01%     
+ Complexity     4979     4978       -1     
============================================
  Files           554      554              
  Lines         14731    14739       +8     
  Branches       1376     1376              
============================================
+ Hits          13453    13459       +6     
  Misses          884      884              
- Partials        394      396       +2     

Impacted Files	Coverage Δ
...metry/sdk/logs/export/BatchLogRecordProcessor.java	91.03% <100.00%> (-0.46%)	⬇️
...telemetry/sdk/trace/export/BatchSpanProcessor.java	94.63% <100.00%> (+0.14%)	⬆️

... and 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[JoeCqupt]

@jkwatson @jack-berg PTAL

[jkwatson]

This removes an existing attribute, correct? I'm worried it will break existing dashboards if we get rid of an attribute. @jack-berg is there a spec for these metrics yet?

[JoeCqupt]

this modification is incompatible。 If you think compatibility is necessary, I can keep the original label and then add new labels

[breedx-splk]

Strongly recommend avoiding using the term "label" to refer to OpenTelemetry attributes. Let's try and be consistent in our naming. 🙏🏻

[jack-berg]

This is being actively discussed in open-telemetry/semantic-conventions#184. In order to reduce churn, I'm inclined to hold any changes to these metrics until a new set of experimental semantic conventions lands.

@JoeCqupt would you be able to comment over on that PR with your use case / requirements?