Container logging OTEL not properly handled.

[Annelotte-Mons]

Version

6.3.x

Expected behavior

When encountering failures / warnings, I would expect the traces to be clearly visible in the application insights with proper linkage. Where I could also filter by instance based on the roleName.

Actual behavior

Ever since Invictus changed to the OTEL versions of the application insights SDK (I think), the logs are not showing up properly in Application Insights. You have to dive into logs and execute complex KQL queries to find individual traces scattered around.

These traces also have the role name of the instance of the container which is default behaviour (so if scales, or swapped out, it becomes a different instance). So trying to filter on just logs from the ImportJob means you have to toggle 50+ roles for example. There is an app setting for this that you can set to your containerAppName to properly group them.

Currently it also seems that I can find no traces or logs from the dashboard and dashboard gateway whatsoever. So, debugging an issue becomes a time consuming job.

Steps to reproduce

No response

Additional context

No response

[stijnmoreels]

Added to the backlog below our general improvement of observability and migration to OpenTelemetry in our import jobs.

[Annelotte-Mons]

fyi : for container Apps using Microsoft.ApplicationInsights.AspNetCore 3.x package we solved the role name issue by setting the appSetting 'OTEL_SERVICE_NAME'

// Commmon app settings
var commonAppSettings = [
{
name: 'ARCUS_HTTP_PORT'
value: '${ingressPort}'
}
{
name: 'APPLICATIONINSIGHTS_CONNECTION_STRING'
value: appInsights.properties.ConnectionString
}
{
name: 'OTEL_SERVICE_NAME'
value: containerAppName
}
]