APM -http.server.request.duration metrics coming as grouped instead of Histogram type

[Maheswari1812]

http.server.request.duration

histogram

Duration of HTTP server requests.

metrics coming as grouped like below

• http.server.request.duration_sum
  http.server.request.duration_count
  http.server.request.duration_max
  http.server.request.duration_bucket
  http.server.request.duration_min
• http.client.request.duration_count similarly...

http.route as well coming as Gain/Vl/* instead of full end point. Any solution for this. 

[Maheswari1812]

Splunk Observability endpoint grouping settings adjustment
http.route coming as A/* - grouped endpoint
but url.path coming as A/B/C/D - full endpoint

how to fix this issue ? Anyone can help?

[ITWhisperer]

What is your search/SPL for this?

[Maheswari1812]

A = data('http.server.request.duration_min', filter=filter('http.route', '/Gain/vl/*')).publish(label='A') -signalflow builder query.

I need histogram function metrics instead of data. Is there any configuration changes required as part of otel instrumentation.

A = histogram('http.server.request.duration', filter=filter('http.route', '/batch-process/iomatch')).min().publish(label='A')

[premadhas]

Hi, @Maheswari1812 You was able to figureout the solution for this? If so can you please share?

I am also experiencing the same issue.
In my case, I want to set up a monitor to track latency for individual endpoints. However, when I search for the metric http.server.request.duration in the Observability Metric Finder, I see it listed (as shown in the screenshot), but when I click on it, it does not show the direct metric data. Instead, it only displays the aggregated metrics such as sum, count, min, max, and bucket, which are derived from http.server.request.duration.

My requirement is to access http.server.request.duration itself so that I can configure a monitor based directly on the latency duration. Could you please guide me on how to access this metric or suggest the correct metric to use for endpoint-level latency monitoring?

[CesarF]

I appreciate this is a (very) late reply! The signalfx exporter breaks down the histogram into their aggregates metrics by default (backward compability).

Use the parameter send_otlp_histograms on the signalfx exporter config and you will get the proper histogram.

https://github.com/open-telemetry/opentelemetry-collector-contrib/tree/main/exporter/signalfxexporte...