[Question] Too much traces /track being sent

[lapa182]

Hello! We recently noticed a spike in our Next.js application of traces coming from /track. I was googling about it and saw this Github issue in another SDK (microsoft/ApplicationInsights-Python#317).

Do you know how I could achieve the same in the Node SDK?

[JacksonWeber]

@lapa182 Can you provide some further details here? Are you seeing duplicated traces? Or do you just want to see fewer traces in general based on some kind of criteria? Also, what version of the SDK are you using/did you recently upgrade before seeing this spike?

[lapa182]

Hi @JacksonWeber, sorry for the delay was trying to get some information as we had to disable AppInsights in our instance because our logging was ramping up our costs.

I got a screenshot of the "issue":

This is a dev instance and you can see that every minute (sometimes less than a minute) we had one track event from v2/track endpoint.

For context, we are using Next.js 14 and we are loading the instrumentation with this code:

	appInsights
		.setup(connectionString)
		.setAutoCollectRequests(true)
		.setAutoCollectPerformance(true, true)
		.setAutoCollectExceptions(true)
		.setAutoCollectDependencies(true)
		.setAutoCollectConsole(true, true)
		.setAutoCollectPreAggregatedMetrics(true)
		.setSendLiveMetrics(false)
		.setInternalLogging(false, true)
		.enableWebInstrumentation(false)
		.start();

We were using this implementation before https://github.com/CMeeg/nextjs-aca/blob/main/src/lib/instrumentation/azure-monitor.ts, which was still in beta and when the SDK came out of beta we decided to update the code (we are on 3.1.0 now). Is there a way to disable capturing v2/track or do you have any other suggestions we could do?

Not sure if it's related as well, but we are using Log Analytics in App Insights.

[hectorhdzg]

@lapa182 I tried to reproduce with no success using Next.js, are you calling setup and start in Application Insights multiple times in your Node.js process?, what you are experiencing is not a common thing, you can always explicitly remove these extra dependencies using HTTP OpenTelemetry Instrumentation that is being used internally to generate the telemetry configuring [ignoreOutgoingRequestHook]
(https://github.com/open-telemetry/opentelemetry-js/tree/main/experimental/packages/opentelemetry-instrumentation-http#http-instrumentation-options), you should be able to add this configuration in 'instrumentationOptions' 'http' configuration if using @azure/monitor-opentelemetry directly.