[Bug]: SPM - Error rate % not showing 4xx client errors.

[mayank-rocketml]

What happened?

4xx errors are not getting displayed in Error rate % (SPM - Monitor Tab).

The reason is that the HTTP status codes in the 4xx range span status is being left unset in case of SpanKind.SERVER by Span Metrics processor (open-telemetry/opentelemetry-collector-contrib#12753).

I think this is a bug.
The Error rate % tab should also consider displaying data for 4xx errors (span_kind="SPAN_KIND_SERVER", status_code="STATUS_CODE_UNSET")

Steps to reproduce

Generate 4xx errors from an application. Traces will appear in Jaeger but the Monitor tab will not show it as an error.

Expected behavior

Error rate % graph in Monitor tab should consider displaying data for 4xx errors.

Relevant log output

No response

Screenshot

No response

Additional context

No response

Jaeger backend version

No response

SDK

No response

Pipeline

No response

Stogage backend

No response

Operating system

No response

Deployment model

No response

Deployment configs

No response

[juferreira]

Hello. Any update for that?
I have the same problem.

[albertteoh]

SPM tab supports querying metrics on span kind; if you're capturing client spans, you should be able to see these error metrics in SPM.

Otherwise, if we're not collecting client spans, could we use collector processors (e.g. attributesprocessor) to set status_code="STATUS_CODE_ERROR" if the http.response.status_code is 4xx?

I think we should try to avoid modifying SPM functionality to support error edge cases where status_code != "STATUS_CODE_ERROR".

[juferreira]

I'm using the last version of helm chart (0.73.1), but I realised that the default version of All-In-One deployment on this chart are 1.51.0. I updated the tag version of the deployment to 1.53.0 and now I have the filter Span Kind and all the erros available.
`
allInOne:

      tag: 1.53.0

      args: 
      - --prometheus.query.support-spanmetrics-connector=true
      - --query.enable-tracing=true

      enabled: true

      extraEnv: 
        - name: METRICS_STORAGE_TYPE
          value: prometheus
        - name: PROMETHEUS_SERVER_URL
          value: http://prometheus.observability.svc.cluster.local:9090
        - name: PROMETHEUS_QUERY_NORMALIZE_CALLS
          value: 'true'
        - name: PROMETHEUS_QUERY_NORMALIZE_DURATION
          value: 'true'

`

And if I search for traces (in the Search tab) using the http.status_code=400 Tag, I can also see all logs with status code 400.

I hope this can help.

[jkowall]

Per @albertteoh I do not think we should support this edge case in SPM.

I think we should try to avoid modifying SPM functionality to support error edge cases where status_code != "STATUS_CODE_ERROR".