Azure Application Insights is not receiving telemetry from local

Question

Azure Application Insights is not receiving telemetry from local

Maria Natalia Benavides Sanabria 10

Azure Application Insights was configured in two backend services running locally, one in Node.js and the other in Python. However, no telemetry is being displayed in Azure Portal. No data appears in tables such as traces, requests, dependencies, or customEvents.

The connection string is configured through environment variables and is successfully validated at application startup. The Application Insights resource is enabled in Azure, and the user has the required access role on the resource.

Environment

ComponentDetailNode.jsv24.13.0----------------Node.jsv24.13.0TypeScript^5.3.2Python3.12.0Operating systemWindows 11 (local)Node.js SDK installed@azure/monitor-opentelemetry@1.16.0Python SDKazure-monitor-opentelemetry## Current implementation

Node.js

Telemetry is initialized at application startup using the following implementation:

import * as appInsights from 'applicationinsights';

export function initTelemetry(connectionString: string | undefined): void {
    if (!connectionString) {
        console.warn('[Telemetry] Connection string is not defined. Telemetry disabled.');
        return;
    }

    appInsights
        .setup(connectionString)
        .setAutoDependencyCorrelation(true)
        .setAutoCollectRequests(true)
        .setAutoCollectPerformance(true)
        .setAutoCollectExceptions(true)
        .setAutoCollectDependencies(true)
        .setAutoCollectConsole(true, true)
        .setUseDiskRetryCaching(true)
        .setSendLiveMetrics(true)
        .start();

    appInsights.defaultClient.trackEvent({
        name: 'AppStartup',
        properties: { 'telemetry.startup': 'true' }
    });
}

Telemetry is initialized before the rest of the application starts.

Python

Azure Monitor is configured at startup using configure_azure_monitor, and custom events are sent through logging.

Environment variable

APPLICATIONINSIGHTS_CONNECTION_STRING=InstrumentationKey=...;IngestionEndpoint=...;...

Current Azure status

Application Insights resource is enabled
Local application is enabled in the resource
User role assigned: Monitoring Metrics Publisher

Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-03-30T20:16:34.27+00:00
Hello Maria Natalia Benavides Sanabria

Thank you for posting your query on Microsoft Q&A portal and sharing the full context and implementation details. Based on your setup (local Node.js + Python apps, Windows 11, OpenTelemetry–based SDKs, validated connection string, but zero data in all Application Insights tables).

You are mixing two different telemetry pipelines that are not compatible as currently implemented

Node.js

You are initializing telemetry using the legacy applicationinsights SDK

At the same time, your environment uses OpenTelemetry-based packages (@azure/monitor-opentelemetry)

Python

Uses Azure Monitor OpenTelemetry Distro (azure-monitor-opentelemetry)

Azure Monitor Application Insights must be instrumented consistently using OpenTelemetry when using the Azure Monitor OpenTelemetry Distro. Mixing legacy SDKs and OpenTelemetry SDKs is not supported and can result in telemetry being dropped silently.

Reference: https://learn.microsoft.com/en-us/azure/azure-monitor/app/opentelemetry-enable?tabs=aspnetcore

Below are the reasons for Telemetry Not Appearing Even Though Setup Looks Correct:

Legacy SDK vs OpenTelemetry conflict (Node.js):

Your Node.js code uses:

import * as appInsights from 'applicationinsights';

However:

applicationinsights v3.x is OpenTelemetry-based

@azure/monitor-opentelemetry is also OpenTelemetry-based

Running both pipelines causes duplicate exporters and span processors

Result: telemetry never reaches ingestion

Microsoft explicitly warns to use only one OpenTelemetry entry point per application.

Unsupported Node.js runtime: You are running node.js v24.13.0 but Azure Monitor OpenTelemetry does not yet support Node.js 24. Supported versions are only those officially supported by OpenTelemetry.

Running on an unsupported runtime results in no exporter initialization, with no visible errors.

Python: traces may exist but not export without active instrumentation

In Python: configure_azure_monitor(connection_string=...)

The Azure Monitor OpenTelemetry Python distro does not automatically instrument everything

A tracer span alone may not appear unless:

Requests / frameworks (Flask, FastAPI, etc.) are instrumented

OR logging is correctly bridged to OpenTelemetry

Reference: https://learn.microsoft.com/en-us/python/api/overview/azure/monitor-opentelemetry-readme?view=azure-python

Local outbound connectivity to ingestion endpoints:

When running locally on Windows:

Telemetry is sent over HTTPS to:

{region}.in.applicationinsights.azure.com

dc.services.visualstudio.com

If outbound HTTPS (443) is blocked by:

Corporate firewall

VPN

Endpoint security software

Telemetry is silently dropped.

Please look into below workaround:

Do not use the applicationinsights SDK.

Use only Azure Monitor OpenTelemetry:
import { useAzureMonitor } from "@azure/monitor-opentelemetry"; useAzureMonitor({ connectionString: process.env.APPLICATIONINSIGHTS_CONNECTION_STRING, });

Remove:

applicationinsights

setup()

defaultClient.trackEvent

Use a supported Node.js version: v18 or v20 (LTS).

Ensure telemetry is actually generated:

Use a supported framework (Flask, FastAPI, requests), or

Create spans after calling configure_azure_monitor().

Use the standard logging module (not print).

Outbound HTTPS (443) must be allowed to:

dc.services.visualstudio.com

{region}.in.applicationinsights.azure.com

Role assigned: Monitoring Metrics Publisher is Sufficient and Not required to send telemetry

Telemetry ingestion does not depend on RBAC.

Thanks,

Suchitra.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-01T19:27:31.22+00:00

Hello Maria Natalia Benavides Sanabria

Could you please let us know if the comment above resolved your issue? If you need any further assistance, feel free to reach out, we're always here to support you.

If you found the comment helpful, please consider clicking "Upvote it."

Thanks,

Suchitra.
Maria Natalia Benavides Sanabria 10 Reputation points

2026-04-10T00:20:24.8333333+00:00

Hello @Suchitra Suregaunkar ,

Thank you for your follow-up.

The issue has not been resolved yet. The telemetry information is still not reaching Azure Application Insights, so no data is being ingested or reflected there.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-10T10:32:56.8033333+00:00
Hello Maria Natalia Benavides Sanabria

Even when the Application Insights SDK is correctly configured and telemetry is being generated locally, ingestion can still fail if there is an issue in the telemetry processing pipeline.

If telemetry does not appear in the Azure portal, the following conditions may apply:

The SDK is correctly configured, but outbound network connectivity to the ingestion endpoint is blocked.

The ingestion endpoint may reject inbound telemetry due to authentication mismatch (for example, if Microsoft Entra ID–based ingestion is enabled on the Application Insights resource).

DNS resolution failures, firewall restrictions, or proxy settings in the local environment may prevent telemetry from reaching the ingestion endpoint.

In such scenarios, telemetry is silently rejected and does not appear in the Azure portal even though the application is generating telemetry data locally.

As a next step, it is recommended to test direct telemetry ingestion using a sample request (via PowerShell or curl) to determine whether the issue is related to SDK instrumentation or network/ingestion endpoint connectivity.

Reference: https://learn.microsoft.com/en-us/troubleshoot/azure/azure-monitor/app-insights/telemetry/investigate-missing-telemetry?tabs=net#steps-that-can-cause-telemetry-to-be-missing

Thanks,
Suchitra.
Maria Natalia Benavides Sanabria 10 Reputation points

2026-04-22T19:20:41.7033333+00:00

Hello @Suchitra Suregaunkar ,

For now, I would like to mention that, when I send the request to the Azure Application Insights endpoint, I receive a successful response with status code 200. In addition, the metrics charts available in the Azure portal do show activity in the service, matching the time at which I performed the test.

However, when I query the logs using KQL, no results are returned. This could indicate that the telemetry is being partially received by the service, at least at the metrics level, but is not being reflected in the queryable logs.

In this regard, I would appreciate it if you could let me know which permissions, configurations, or validations should be reviewed within the corporate environment or Azure service, especially regarding ingestion, retention, the associated workspace, log enablement, and the required access levels to query the information.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-23T09:20:01.3433333+00:00

Hello Maria Natalia Benavides Sanabria

Sure, let me check and will update you.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-28T11:30:42.68+00:00
Hello Maria Natalia Benavides Sanabria

If metrics charts are showing activity while KQL queries return no log data, it confirms that telemetry is reaching Azure, but it is not being ingested into the Log Analytics–backed tables (AppTraces, Requests, Dependencies, etc.). This points to a workspace or log ingestion configuration issue, not a connectivity problem.

Below are the specific validations to review, in order:

Verify Application Insights workspace-based configuration :

Please confirm that the Application Insights resource is workspace-based and correctly linked:

Go to Application Insights → Properties

Check Workspace:

It must be set

Note the exact Log Analytics workspace name

Run all KQL queries in that same workspace, not a different one.

Metrics do not depend on the workspace, but logs do, which is why metrics can appear while KQL shows nothing.

Confirm log ingestion is enabled (not just metrics)

In Application Insights → Diagnostic settings:

Ensure diagnostics are enabled

Confirm that logs are being sent to:

Log Analytics workspace

At least one of the following categories should be enabled:

AppTraces

AppRequests

AppDependencies

AppExceptions

If no diagnostic setting exists, logs will not be queryable, even though metrics continue to work.

Validate retention settings: Check retention at two levels:

Application Insights

AI resource retention only affects metrics and availability

Log Analytics Workspace

Go to Workspace → Usage and estimated costs → Data retention

Ensure retention is > 0 days

If set very low (or data already aged out), KQL will return no rows

Access required to query logs (RBAC): To query logs, the user must have one of these roles on the Log Analytics workspace:

Log Analytics Reader

Monitoring Reader

Contributor (or higher)

Note: RBAC is not required for telemetry ingestion, but is required to query logs. Metrics visibility alone does not guarantee log query access.

Why metrics can appear when logs do not:

This behavior is expected in certain scenarios because:

Metrics are ingested through a lightweight, separate pipeline

Logs require:

Workspace association

Log ingestion enabled

Compatible SDK pipeline (OpenTelemetry vs legacy)

When log ingestion fails or is misconfigured, metrics still flow

This explains why you see HTTP 200 ingestion responses, Metrics activity and Empty KQL tables.

Once workspace and diagnostics are confirmed, run this in the linked workspace:

union AppTraces, AppRequests, AppDependencies, AppExceptions | order by TimeGenerated desc | take 20

If this still returns no data, then log export from the SDK is not occurring (pipeline/runtime issue).

Metrics working but logs empty indicates workspace/log ingestion configuration or access, not connectivity or instrumentation key issues.

Thanks,
Suchitra.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-29T17:22:28.2+00:00

Hello Maria Natalia Benavides Sanabria

Could you please let us know if the comment above resolved your issue? If you need any further assistance, feel free to reach out—we're always here to support you.

If you found the comment helpful, please consider clicking "Upvote it."

Thanks,

Suchitra.
Maria Natalia Benavides Sanabria 10 Reputation points

2026-06-11T00:37:35.6033333+00:00
Hello @Suchitra Suregaunkar y @Sina Salam
Thank you for the detailed checklist. Here's everything we've already verified, point by point, so you have full context:

Workspace-based configuration (confirmed): In Application Insights -> Properties, the resource is workspace-based and linked to its corresponding Log Analytics workspace. We run the KQL queries against that same linked workspace.

Correct scope / table names (verified): We noticed the table names differ by scope and tested both:

From Application Insights -> Logs we used traces, requests, dependencies, exceptions.

From Log Analytics workspace -> Logs we used AppTraces, AppRequests, AppDependencies, AppExceptions.

Using AppTraces in the AI scope returns "Failed to resolve", and traces in the workspace scope returns empty — so we confirmed it's not a table-name issue; we queried each table in its correct scope.

Ingestion (confirmed working).

Overview shows real activity: real server requests.

A manual test POST to the ingestion endpoint returned HTTP 200.

The network path is clean: no proxy, valid Microsoft certificate. So connectivity and the instrumentation key are fine.

Instrumentation/SDK (verified): Telemetry is emitted via the OpenTelemetry SDK (configure_azure_monitor in Python, useAzureMonitor in Node), and custom events via logger.info(..., extra={...}), which should land in traces / AppTraces. The code is confirmed correct, with no pending changes.

Time range (ruled out): We set the time range wide / "set in query" so the UI time picker isn't filtering results out.

Diagnostic settings: Since this is a workspace-based resource using the OpenTelemetry connection string, our understanding is that the AppTraces/AppRequests/...tables populate automatically without a separate diagnostic setting. Please confirm if that's not the case here.
RBAC (the open item): The user running the KQL currently does not have the Log Analytics Reader (or Monitoring Reader) role on the workspace. The symptom matches exactly: Metrics/Overview show data, but the KQL returns 0 rows with no error. We couldn't validate your union AppTraces, AppRequests,AppDependencies, AppExceptions | order by TimeGenerated desc | take 20 query, because without read access we can't tell whether the tables are empty or simply not readable.

That said, the goal is not for each user to need specific permissions for their activity to be recorded. What we want is for any user who enters the application to report their telemetry automatically, without depending on individual roles. So we'd like to confirm whether the lack of this role affects only the reading/querying of the logs (not the capture), or whether it also affects whether the telemetry gets recorded at all.

Ingestion returns HTTP 200, Metrics/Overview show activity, but every KQL query returns 0 rows with no error, across both scopes and with a wide time range.

Our question and, based on the above, what could the cause be? Is the empty KQL due to the missing Log Analytics Reader role on the workspace (i.e., the logs are there but can't be read), or could it still be an ingestion/pipeline issue despite the workspace-based + OpenTelemetry setup? What would you recommend as the next step to confirm and resolve it?

1 answer

Your answer

Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-01T19:27:31.22+00:00

Hello Maria Natalia Benavides Sanabria

Could you please let us know if the comment above resolved your issue? If you need any further assistance, feel free to reach out, we're always here to support you.

If you found the comment helpful, please consider clicking "Upvote it."

Thanks,

Suchitra.
Maria Natalia Benavides Sanabria 10 Reputation points

2026-04-10T00:20:24.8333333+00:00

Hello @Suchitra Suregaunkar ,

Thank you for your follow-up.

The issue has not been resolved yet. The telemetry information is still not reaching Azure Application Insights, so no data is being ingested or reflected there.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-10T10:32:56.8033333+00:00

Hello Maria Natalia Benavides Sanabria

Even when the Application Insights SDK is correctly configured and telemetry is being generated locally, ingestion can still fail if there is an issue in the telemetry processing pipeline.

If telemetry does not appear in the Azure portal, the following conditions may apply:

The SDK is correctly configured, but outbound network connectivity to the ingestion endpoint is blocked.

The ingestion endpoint may reject inbound telemetry due to authentication mismatch (for example, if Microsoft Entra ID–based ingestion is enabled on the Application Insights resource).

DNS resolution failures, firewall restrictions, or proxy settings in the local environment may prevent telemetry from reaching the ingestion endpoint.

In such scenarios, telemetry is silently rejected and does not appear in the Azure portal even though the application is generating telemetry data locally.

As a next step, it is recommended to test direct telemetry ingestion using a sample request (via PowerShell or curl) to determine whether the issue is related to SDK instrumentation or network/ingestion endpoint connectivity.

Reference: https://learn.microsoft.com/en-us/troubleshoot/azure/azure-monitor/app-insights/telemetry/investigate-missing-telemetry?tabs=net#steps-that-can-cause-telemetry-to-be-missing

Thanks,
Suchitra.
Maria Natalia Benavides Sanabria 10 Reputation points

2026-04-22T19:20:41.7033333+00:00

Hello @Suchitra Suregaunkar ,

For now, I would like to mention that, when I send the request to the Azure Application Insights endpoint, I receive a successful response with status code 200. In addition, the metrics charts available in the Azure portal do show activity in the service, matching the time at which I performed the test.

However, when I query the logs using KQL, no results are returned. This could indicate that the telemetry is being partially received by the service, at least at the metrics level, but is not being reflected in the queryable logs.

In this regard, I would appreciate it if you could let me know which permissions, configurations, or validations should be reviewed within the corporate environment or Azure service, especially regarding ingestion, retention, the associated workspace, log enablement, and the required access levels to query the information.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-23T09:20:01.3433333+00:00

Hello Maria Natalia Benavides Sanabria

Sure, let me check and will update you.
Suchitra Suregaunkar 16,700 Reputation points Microsoft External Staff Moderator

2026-04-29T17:22:28.2+00:00

Hello Maria Natalia Benavides Sanabria

Could you please let us know if the comment above resolved your issue? If you need any further assistance, feel free to reach out—we're always here to support you.

If you found the comment helpful, please consider clicking "Upvote it."

Thanks,

Suchitra.

Answer 1

Sina Salam 31,456 Volunteer Moderator

Hello Maria Natalia Benavides Sanabria,

Welcome to the Microsoft Q&A and thank you for posting your questions here.

I understand that your Azure Application Insights is not receiving telemetry from local.

To resolve all the issues. Use LTS Node.js and initialize Azure Monitor OpenTelemetry first (useAzureMonitor before any imports), configure Python with configure_azure_monitor, ensure Entra ID ingestion isn’t blocking key-based telemetry, then query the linked workspace using AppTraces/AppRequests/... tables with Monitoring Reader (or equivalent) permissions. For more steps by steps use the following link: https://learn.microsoft.com/en-us/javascript/api/overview/azure/monitor-opentelemetry-readme?view=azure-node-latest, https://www.npmjs.com/package/@azure/monitor-opentelemetry, https://learn.microsoft.com/en-us/python/api/overview/azure/monitor-opentelemetry-readme?view=azure-python, https://learn.microsoft.com/en-us/troubleshoot/azure/azure-monitor/app-insights/telemetry/investigate-missing-telemetry, https://learn.microsoft.com/en-us/azure/azure-monitor/app/create-workspace-resource, https://docs.azure.cn/en-us/azure-monitor/reference/tables/microsoft-insights_components, https://learn.microsoft.com/en-us/azure/azure-monitor/fundamentals/roles-permissions-security, https://learn.microsoft.com/en-us/azure/azure-monitor/logs/manage-access

I hope this is helpful! Do not hesitate to let me know if you have any other questions or clarifications.

Please don't forget to close up the thread here by upvoting and accept it as an answer if it is helpful.

Sina Salam 31,456 Reputation points Volunteer Moderator

2026-06-11T13:35:05.9666667+00:00
Hello Maria Natalia Benavides Sanabria,

Thank you for your feedback and detail information.

Now, we can be certain that Telemetry is being ingested or at least reaching Application Insights, but the current Azure user cannot validate it in KQL because of lack of log-read permission on the linked Log Analytics workspace.

You can fix as below:

Assign one of these roles to Azure user or, preferably, to an Entra ID group:

Monitoring Reader, or

Log Analytics Reader, or

Log Analytics Data Reader / granular RBAC if table-level restrictions are required.

Assign it on the linked Log Analytics workspace, not only on the Application Insights resource.

Do not use Monitoring Metrics Publisher for this. Monitoring Reader is documented as allowing users to search Log Analytics workspace data, retrieve schemas, execute log queries, and access Application Insights data. - https://learn.microsoft.com/en-us/azure/azure-monitor/fundamentals/roles-permissions-security, https://learn.microsoft.com/en-us/azure/azure-monitor/logs/manage-access

Secondly, wait a few minutes, then sign out/sign in or refresh the Azure Portal session.

Thirdly, use the below from linked log analytic workspace.

union isfuzzy=true AppRequests, AppTraces, AppDependencies, AppExceptions | where TimeGenerated > ago(24h) | project TimeGenerated, Type, AppRoleName, AppRoleInstance, OperationName, Message, Properties, SeverityLevel, _ResourceId, IKey | order by TimeGenerated desc | take 100

If this returns rows after the role assignment, the root cause is confirmed: RBAC was hiding the data from the querying user.

Fourthly, if there is still zero rows after correct RBAC - Only then investigate SDK/pipeline. The minimum checks are: Node.js. Use the OpenTelemetry distro correctly:

import { useAzureMonitor, AzureMonitorOpenTelemetryOptions } from "@azure/monitor-opentelemetry"; const options: AzureMonitorOpenTelemetryOptions = { azureMonitorExporterOptions: { connectionString: process.env["APPLICATIONINSIGHTS_CONNECTION_STRING"], }, }; useAzureMonitor(options);

useAzureMonitor() must run before importing instrumented modules. Microsoft warns that telemetry loss can occur if other libraries are imported first. - https://learn.microsoft.com/en-us/javascript/api/overview/azure/monitor-opentelemetry-readme?view=azure-node-latest Then use:

from azure.monitor.opentelemetry import configure_azure_monitor configure_azure_monitor()

The connection string may be set through APPLICATIONINSIGHTS_CONNECTION_STRING; Microsoft documents the connection string as the target for Application Insights telemetry. - https://learn.microsoft.com/en-us/azure/azure-monitor/app/opentelemetry-configuration, and https://docs.azure.cn/en-us/azure-monitor/app/connection-strings

Success.