Azure Data Explorer Exporter - Missing Exception Fields from OTelLogs and OTelTraces #26496

akakarikos · 2023-09-07T07:43:51Z

Component(s)

exporter/azuredataexplorer

What happened?

Description

When there is an unhandled exception in the application and captured by an instrumentation library, the exception details such as exception message and exception stack trace do not populate in the Azure Data Explorer tables logs and traces.

To give you an example:

See below a comparison between a trace record from the OTel Collector console and the ADX Traces table. You will notice that the Status Message field is missing from the ADX trace.

OTel Collector Console

ScopeSpans #1
ScopeSpans SchemaURL:
InstrumentationScope OpenTelemetry.Instrumentation.AspNetCore 1.0.0.0
Span #0
    Trace ID       : 96eb74d68c214c468798fd3462bc46aa
    Parent ID      : 102608693b89e619
    ID             : 20af91f77a0fb7a5
    Name           : endpoint_name
    Kind           : Server
    Start time     : 2023-09-06 16:14:17.0062245 +0000 UTC
    End time       : 2023-09-06 16:14:18.9010324 +0000 UTC
    Status code    : Error
    Status message : error message displayed here
Attributes:
     -> net.host.name: Str(127.0.0.1)
     -> net.host.port: Int(8080)
     -> http.method: Str(POST)
     -> http.scheme: Str(http)
     -> http.target: Str(/endpoint_name)
     -> http.url: Str(http://127.0.0.1:8080/endpoint_name)
     -> http.flavor: Str(1.1)
     -> http.user_agent: Str(dapr-sdk-dotnet/v1.11.0+787cba6f4f4d28bd8ec09fcef9c854ef64d4361b)
     -> http.route: Str(endpoint_name)
     -> http.status_code: Int(500)
     -> ThreadId: Str(17)
     -> ReferenceId: Str(9a74cefd-62ab-4a4f-8572-b53090df083f)
     -> TransactionId: Str()

ADX Traces Table

"TraceID": 96eb74d68c214c468798fd3462bc46aa,
"SpanID": 20af91f77a0fb7a5,
"ParentID": 102608693b89e619,
"SpanName": endpoint_name,
"SpanStatus": STATUS_CODE_ERROR,
"SpanKind": SPAN_KIND_SERVER,
"StartTime": 2023-09-06T16:14:17.0062245Z,
"EndTime": 2023-09-06T16:14:18.9010324Z,
"ResourceAttributes": {
	"service.name": "service_name",
	"service.version": "9.0.0.0",
	"service.instance.id": "4978bae1-a67c-4b76-9a60-80af647cc2c4",
	"telemetry.sdk.name": "opentelemetry",
	"telemetry.sdk.language": "dotnet",
	"telemetry.sdk.version": "1.6.0-rc.1"
},
"TraceAttributes": {
	"http.url": "[http://127.0.0.1:8080/endpoint_name"](http://127.0.0.1:8080/endpoint_name%22),
	"http.target": "endpoint_name",
	"net.host.port": 8080,
	"ThreadId": "17",
	"http.user_agent": "dapr-sdk-dotnet/v1.11.0+787cba6f4f4d28bd8ec09fcef9c854ef64d4361b",
	"scope.name": "OpenTelemetry.Instrumentation.AspNetCore",
	"http.route": "endpoint_name",
	"ReferenceId": "9a74cefd-62ab-4a4f-8572-b53090df083f",
	"http.scheme": "http",
	"http.flavor": "1.1",
	"scope.version": "1.0.0.0",
	"TransactionId": "",
	"http.status_code": 500,
	"net.host.name": "127.0.0.1",
	"http.method": "POST"
},
"Events": [],
"Links": []

See below a comparison between a log record from the Application Console and the ADX Logs table. You will notice that the LogRecord.Exception field is missing from the ADX log record, so we cannot retrieve the exact exception details and stack trace.

Application Console

LogRecord.Timestamp:               2023-09-06T16:14:18.8893862Z
LogRecord.TraceId:                 96eb74d68c214c468798fd3462bc46aa
LogRecord.SpanId:                  20af91f77a0fb7a5
LogRecord.TraceFlags:              Recorded
LogRecord.CategoryName:            Microsoft.AspNetCore.Server.Kestrel
LogRecord.Severity:                Error
LogRecord.SeverityText:            Error
LogRecord.Body:                    exception message displayed here
LogRecord.Attributes (Key:Value):
    ConnectionId: 0HMTEUH40C96J
    TraceIdentifier: 0HMTEUH40C96J:00000004
    OriginalFormat exception message displayed here
    ThreadId: 17
    ReferenceId: 9a74cefd-62ab-4a4f-8572-b53090df083f
    TransactionId:
LogRecord.EventId:                 13
LogRecord.EventName:               ApplicationError
LogRecord.Exception:               specific exception message and stacktrace displayed here
LogRecord.ScopeValues (Key:Value):
[Scope.0]:SpanId: 20af91f77a0fb7a5
[Scope.0]:TraceId: 96eb74d68c214c468798fd3462bc46aa
[Scope.0]:ParentId: 102608693b89e619
[Scope.1]:ConnectionId: 0HMTEUH40C96J
[Scope.2]:RequestId: 0HMTEUH40C96J:00000004
[Scope.2]:RequestPath: endpoint_name

ADX Logs Table

"Timestamp": 2023-09-06T16:14:18.8893862Z,
"ObservedTimestamp": 2023-09-06T16:14:18.8893862Z,
"TraceID": 96eb74d68c214c468798fd3462bc46aa,
"SpanID": 20af91f77a0fb7a5,
"SeverityText": Error,
"SeverityNumber": 17,
"Body": exception message displayed here,
"ResourceAttributes": {
	"telemetry.sdk.name": "opentelemetry",
	"telemetry.sdk.language": "dotnet",
	"telemetry.sdk.version": "1.6.0-rc.1",
	"service.name": "service_name",
	"service.version": "9.0.0.0",
	"service.instance.id": "c91fc583-cbcf-4845-aac4-fe5eb9024d77"
},
"LogsAttributes": {
	"SpanId": "20af91f77a0fb7a5",
	"ParentId": "102608693b89e619",
	"RequestId": "0HMTEUH40C96J:00000004",
	"ReferenceId": "9a74cefd-62ab-4a4f-8572-b53090df083f",
	"ThreadId": 17,
	"TraceId": "96eb74d68c214c468798fd3462bc46aa",
	"RequestPath": "",
	"ConnectionId": "0HMTEUH40C96J",
	"TransactionId": "",
	"TraceIdentifier": "0HMTEUH40C96J:00000004"
}

Steps to Reproduce

Force an application to throw an unhandled exception that will be caught by the OTel collector and exported to the Azure Data Explorer.

Expected Result

The exception details such as exception message, exception stack trace, and exception source are successfully exported in the OTelLogs and OTelTraces ADX tables.

Actual Result

The exception details such as exception message, stack trace, and source are not included in the OTelLogs and OTelTraces ADX table records.

Collector version

0.83.0

Environment information

Environment

Docker image deployed in AKS.
OS: linux amd64

OpenTelemetry Collector configuration

apiVersion: v1
kind: ConfigMap
metadata:
  name: otel-collector-config
data:
  config.yaml: |-
    receivers:
      zipkin:
        endpoint: localhost:9411
      otlp:
        protocols:
          grpc:
            endpoint: localhost:4317
          http:
            endpoint: localhost:4318
            
    processors:
      memory_limiter:
        check_interval: 1s
        limit_percentage: 80
        spike_limit_percentage: 30

      batch:
        send_batch_size: 50
    
    exporters:
      logging:
        verbosity: detailed
    
      azuremonitor:
        instrumentation_key: 
        spaneventsenabled: true

      azuredataexplorer:
        cluster_uri: 
        application_id: 
        application_key: 
        tenant_id:
        db_name: 
        metrics_table_name: 
        logs_table_name: 
        traces_table_name: 
        ingestion_type : 'managed'
    
    service:
      pipelines:
        traces:
          receivers: [zipkin,otlp]
          processors: [memory_limiter,batch]
          exporters: [azuredataexplorer,azuremonitor,logging]
        logs:
          receivers: [otlp]
          processors: [memory_limiter,batch]
          exporters: [azuredataexplorer,azuremonitor,logging]

Log output

No response

Additional context

No response

The text was updated successfully, but these errors were encountered:

github-actions · 2023-09-07T07:45:18Z

Pinging code owners:

exporter/azuredataexplorer: @asaharn @ag-ramachandran

See Adding Labels via Comments if you do not have permissions to add labels yourself.

ag-ramachandran · 2023-09-07T09:42:35Z

Hello @akakarikos, Thanks for reporting. Will have a look and provide details about the missing attributes.

cijothomas · 2023-09-07T16:17:07Z

If you are missing Exception, then its likely due to not being exported in the OTLP Exporter itself:
https://github.com/open-telemetry/opentelemetry-dotnet/blob/main/src/OpenTelemetry.Exporter.OpenTelemetryProtocol/CHANGELOG.md#160-rc1

You can verify that by using "OTel Collector Console" output, which would also be missing the Exception from Logs.

akakarikos · 2023-09-07T17:17:04Z

Thanks a lot for the info @cijothomas that makes sense! You are right, the LogRecord.Exception is missing from the OTel collector console.

However, the LogRecord.Exception is there at the application console. I don't know if this related or not.

Also there is the issue with the traces missing the Status Message.

Apologies in advance if I am missing something.

cijothomas · 2023-09-07T17:31:09Z

Thanks a lot for the info @cijothomas that makes sense! You are right, the LogRecord.Exception is missing from the OTel collector console.

However, the LogRecord.Exception is there at the application console. I don't know if this related or not.

The ConsoleExporter still exports Exception, but OTLP Exporter stopped doing it for the 1.6 stable. (Hopefully it'll be back in the 1.7.*)

cijothomas · 2023-09-07T17:33:51Z

Also there is the issue with the traces missing the Status Message.

That looks like a bug in the ADX exporter, it does not seem to do anything with StatusMessage. Will let the owners triage that.

ag-ramachandran · 2023-09-08T10:12:47Z

Hello @akakarikos , Here is an update

We're in the process of adding Status Code and Status Message attributes. The PR is here, will perform some tests track this for closure

https://github.com/ag-ramachandran/opentelemetry-collector-contrib/tree/bugfix/26496105-StatusMessage

cijothomas · 2023-09-08T15:40:15Z

@ag-ramachandran Thanks. The span's statusmessage is a single string message, not attributes.
https://github.com/open-telemetry/opentelemetry-proto/blob/main/opentelemetry/proto/trace/v1/trace.proto#L266

ag-ramachandran · 2023-09-09T14:29:22Z

Hello @cijothomas, You're right. Unfortunately we already had SpanStatus (which housed the StatusCode). Ideally this field should house the Code and Message as a KV or a JSON (in Kusto we could have a dynamic field, like the one we have for Link or Event attributes)
The challenge is the following, if we change this as SpanStatus and existing customers use this, there is a likely chance the field will have incompatibility (change in Data type from string to dynamic/json)

We could say, we can add a field called SpanStatusMessage and alter that as a string in the table. The only thought we had has, in case there is some new attribute added the table could go wider again, where as if was a KV/JSON we could just add an attribute in

While some of it is Kusto/ADX specific, this was really the rationale. What are your thoughts :)

cc: @akakarikos , @asaharn thoughts

BR,Ram

cijothomas · 2023-09-11T14:48:10Z

We could say, we can add a field called SpanStatusMessage and alter that as a string in the table. The only thought we had has, in case there is some new attribute added the table could go wider again

I'd recommend this option. OpenTelemetry Span hasn't added new top-level fields to Span in years, so this should not be a concern.

OR you can chose to make a breaking change and fix it. Agree it'll break users, but I am not sure if this component is declared Stable, so it's generally okay to make breaking changes.

…es (#26682) **Description:** <Describe what has changed.>  Added an optional column in the exported trace data to store the status code and message as a dynamic field. **Link to tracking Issue:** <Issue number if applicable> [#26496](#26496) **Testing:** <Describe what testing was performed and which tests were added.> Performed E2E ingestion tests and added Test Cases for new fields. **Documentation:** <Describe the documentation added.> --------- Co-authored-by: Ramachandran A G <ramacg@microsoft.com> Co-authored-by: Ramachandran A G <106139410+ag-ramachandran@users.noreply.github.com>

ag-ramachandran · 2023-10-13T11:41:28Z

Hello @akakarikos , Is this something you can test and close

akakarikos · 2023-10-13T13:36:08Z

Hello @ag-ramachandran and thanks for all the remediate actions! Looks good for now, If anything comes up, I will reopen the case.

…es (open-telemetry#26682) **Description:** <Describe what has changed.>  Added an optional column in the exported trace data to store the status code and message as a dynamic field. **Link to tracking Issue:** <Issue number if applicable> [open-telemetry#26496](open-telemetry#26496) **Testing:** <Describe what testing was performed and which tests were added.> Performed E2E ingestion tests and added Test Cases for new fields. **Documentation:** <Describe the documentation added.> --------- Co-authored-by: Ramachandran A G <ramacg@microsoft.com> Co-authored-by: Ramachandran A G <106139410+ag-ramachandran@users.noreply.github.com>

akakarikos added bug Something isn't working needs triage New item requiring triage labels Sep 7, 2023

github-actions bot added the exporter/azuredataexplorer label Sep 7, 2023

atoulme removed the needs triage New item requiring triage label Sep 9, 2023

github-actions bot mentioned this issue Sep 12, 2023

Weekly Report: 2023-09-05 - 2023-09-12 kevinslin/opentelemetry-collector-contrib#23

Open

This was referenced Sep 12, 2023

Weekly Report: 2023-09-05 - 2023-09-12 kevinslin/opentelemetry-collector-contrib#24

Open

Weekly Report: 2023-09-05 - 2023-09-12 kevinslin/opentelemetry-collector-contrib#25

Open

asaharn mentioned this issue Sep 14, 2023

[exporter/azuredataexplorer] Status Attributes added to exported traces #26682

Merged

akakarikos closed this as completed Oct 13, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Azure Data Explorer Exporter - Missing Exception Fields from OTelLogs and OTelTraces #26496

Azure Data Explorer Exporter - Missing Exception Fields from OTelLogs and OTelTraces #26496

akakarikos commented Sep 7, 2023

github-actions bot commented Sep 7, 2023

ag-ramachandran commented Sep 7, 2023

cijothomas commented Sep 7, 2023

akakarikos commented Sep 7, 2023

cijothomas commented Sep 7, 2023

cijothomas commented Sep 7, 2023

ag-ramachandran commented Sep 8, 2023 •

edited

Loading

cijothomas commented Sep 8, 2023

ag-ramachandran commented Sep 9, 2023

cijothomas commented Sep 11, 2023

ag-ramachandran commented Oct 13, 2023

akakarikos commented Oct 13, 2023

Azure Data Explorer Exporter - Missing Exception Fields from OTelLogs and OTelTraces #26496

Azure Data Explorer Exporter - Missing Exception Fields from OTelLogs and OTelTraces #26496

Comments

akakarikos commented Sep 7, 2023

Component(s)

What happened?

Description

Steps to Reproduce

Expected Result

Actual Result

Collector version

Environment information

Environment

OpenTelemetry Collector configuration

Log output

Additional context

github-actions bot commented Sep 7, 2023

ag-ramachandran commented Sep 7, 2023

cijothomas commented Sep 7, 2023

akakarikos commented Sep 7, 2023

cijothomas commented Sep 7, 2023

cijothomas commented Sep 7, 2023

ag-ramachandran commented Sep 8, 2023 • edited Loading

cijothomas commented Sep 8, 2023

ag-ramachandran commented Sep 9, 2023

cijothomas commented Sep 11, 2023

ag-ramachandran commented Oct 13, 2023

akakarikos commented Oct 13, 2023

ag-ramachandran commented Sep 8, 2023 •

edited

Loading