Application Insights Integration¶

Azure App Service allows for deeper application-level monitoring through integration with Application Insights. This provides detailed telemetry on requests, dependencies, exceptions, and overall application performance.

Data Flow Diagram¶

graph TD
    App[Application Code] -->|Instrumentation| AI[Application Insights SDK / Agent]
    AI -->|HTTPS| AIResource[Application Insights Resource]
    AIResource -->|Analysis| Logs[Log Analytics Workspace]
    AIResource -->|Visualization| Dashboards[Azure Dashboards / Metrics Explorer]

Instrumentation Methods¶

There are two primary ways to integrate Application Insights with Azure App Service:

Auto-instrumentation (Codeless): No code changes required. Enabled through App Service app settings. Ideal for quickly adding monitoring to existing apps.
SDK Setup (Code-based): Requires adding the Application Insights SDK to your project code. This method provides more control and allows for custom telemetry.

Configuration Examples¶

Enabling Auto-instrumentation via CLI¶

Use the az webapp config appsettings set command to configure Application Insights for an existing app.

az webapp config appsettings set \
    --resource-group "my-resource-group" \
    --name "my-app-service" \
    --settings "APPLICATIONINSIGHTS_CONNECTION_STRING=InstrumentationKey=00000000-0000-0000-0000-000000000000;IngestionEndpoint=https://centralus-0.in.applicationinsights.azure.com/" \
    "ApplicationInsightsAgent_EXTENSION_VERSION=~2" \
    "XDT_MicrosoftApplicationInsights_Mode=recommended" \
    "XDT_MicrosoftApplicationInsights_PreemptSdk=1"

KQL Query Examples¶

Slowest Requests¶

Identify the top 10 slowest requests in your application to focus optimization efforts.

requests
| where success == true
| order by duration desc
| take 10
| project timestamp, name, duration, success

Dependency Failures¶

Monitor external dependency calls that are failing, such as database or API calls.

dependencies
| where success == false
| summarize count() by name, data
| order by count_ desc

Application Exceptions¶

List the most common exceptions occurring in your application code.

exceptions
| summarize count() by problemId, outerMessage
| order by count_ desc

Correlate Slow Requests With Dependencies¶

Use operation correlation to verify whether a slow request is caused by external services such as SQL, Redis, or HTTP APIs.

requests
| where timestamp > ago(1h)
| where duration > 2s
| project operation_Id, RequestName=name, RequestDuration=duration, resultCode, cloud_RoleName
| join kind=leftouter (
    dependencies
    | where timestamp > ago(1h)
    | project operation_Id, DependencyName=name, DependencyType=type, DependencyDuration=duration, DependencySuccess=success
) on operation_Id
| order by RequestDuration desc

Review Live Failure Trend¶

union requests, dependencies, exceptions
| where timestamp > ago(30m)
| summarize Count=count() by itemType, success, bin(timestamp, 5m)
| order by timestamp asc

Sample output:

timestamp                  itemType      success  Count
-------------------------  ------------  -------  -----
2026-04-06T00:40:00Z       request       true     1580
2026-04-06T00:40:00Z       dependency    false    24
2026-04-06T00:40:00Z       exception              7

Recommended Integration Pattern¶

For App Service, use Application Insights for application behavior, and use App Service platform logs for hosting behavior. The combination answers two different operational questions:

Did the app fail?
- Requests
- Dependencies
- Exceptions
- Traces
Did the platform contribute to the failure?
- App restart events
- Console logs
- HTTP platform logs

This separation helps during incidents. A spike in failed requests with no platform anomalies usually points to code, dependency, or configuration problems.

CLI Configuration Workflow¶

1. Create an Application Insights resource¶

az monitor app-insights component create \
    --app "my-app-insights" \
    --location "koreacentral" \
    --resource-group "my-resource-group" \
    --workspace "/subscriptions/<subscription-id>/resourceGroups/my-resource-group/providers/Microsoft.OperationalInsights/workspaces/law-monitoring-prod" \
    --application-type "web"

Sample output:

{
  "appId": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "applicationType": "web",
  "connectionString": "<connection-string>",
  "name": "my-app-insights"
}

2. Retrieve the connection string¶

az monitor app-insights component show \
    --app "my-app-insights" \
    --resource-group "my-resource-group" \
    --query "connectionString" \
    --output tsv

Sample output:

InstrumentationKey=00000000-0000-0000-0000-000000000000;IngestionEndpoint=https://koreacentral-0.in.applicationinsights.azure.com/;ApplicationId=a1b2c3d4-e5f6-7890-abcd-ef1234567890

3. Validate that settings were applied to the app¶

az webapp config appsettings list \
    --resource-group "my-resource-group" \
    --name "my-app-service" \
    --query "[?name=='APPLICATIONINSIGHTS_CONNECTION_STRING' || name=='ApplicationInsightsAgent_EXTENSION_VERSION']"

Sample output:

[
  {
    "name": "APPLICATIONINSIGHTS_CONNECTION_STRING",
    "value": "InstrumentationKey=00000000-0000-0000-0000-000000000000;IngestionEndpoint=https://koreacentral-0.in.applicationinsights.azure.com/"
  },
  {
    "name": "ApplicationInsightsAgent_EXTENSION_VERSION",
    "value": "~2"
  }
]

Validation Queries After Enablement¶

Run these queries 5 to 15 minutes after deployment or configuration changes.

Confirm request ingestion¶

requests
| where timestamp > ago(15m)
| summarize RequestCount=count(), FailureCount=countif(success == false) by cloud_RoleName

Confirm trace ingestion from application logs¶

traces
| where timestamp > ago(15m)
| summarize TraceCount=count() by severityLevel, cloud_RoleName
| order by severityLevel asc

Confirm exception details include operation correlation¶

exceptions
| where timestamp > ago(1h)
| project timestamp, operation_Id, type, outerMessage, cloud_RoleName
| order by timestamp desc

Practical Alert Examples¶

Scheduled query alert for dependency failure burst¶

az monitor scheduled-query create \
    --name "appsvc-dependency-failures" \
    --resource-group "my-resource-group" \
    --scopes "/subscriptions/<subscription-id>/resourceGroups/my-resource-group/providers/Microsoft.OperationalInsights/workspaces/law-monitoring-prod" \
    --condition "count 'dependencies | where timestamp > ago(5m) | where cloud_RoleName == \"my-app-service\" and success == false' > 25" \
    --description "Dependency failures exceeded threshold for App Service application" \
    --evaluation-frequency "5m" \
    --window-size "5m" \
    --severity 2 \
    --action-groups "/subscriptions/<subscription-id>/resourceGroups/my-resource-group/providers/Microsoft.Insights/actionGroups/ag-app-oncall"

Scheduled query alert for unhandled exceptions¶

az monitor scheduled-query create \
    --name "appsvc-unhandled-exceptions" \
    --resource-group "my-resource-group" \
    --scopes "/subscriptions/<subscription-id>/resourceGroups/my-resource-group/providers/Microsoft.OperationalInsights/workspaces/law-monitoring-prod" \
    --condition "count 'exceptions | where timestamp > ago(10m) | where cloud_RoleName == \"my-app-service\"' > 10" \
    --description "Unhandled exception volume is above normal baseline" \
    --evaluation-frequency "5m" \
    --window-size "10m" \
    --severity 3 \
    --action-groups "/subscriptions/<subscription-id>/resourceGroups/my-resource-group/providers/Microsoft.Insights/actionGroups/ag-app-oncall"

What Good Telemetry Looks Like¶

A well-integrated App Service application normally shows:

Requests with meaningful operation names
Dependencies with target, result, and duration populated
Exceptions linked to the failing request via operation_Id
Traces with enough context to explain the business step or tenant flow
Stable cloud_RoleName values so multi-service queries remain easy to group

Common Pitfalls¶

Enabling codeless instrumentation but not validating that telemetry actually arrives
Treating Application Insights as a replacement for platform logs
Not correlating slow requests with dependency latency
Emitting overly verbose traces and increasing ingestion cost without improving triage quality

Workbook Suggestions¶

Create one Application Insights workbook section for each of these views:

Request rate, failure rate, and duration trend
Top failing dependencies by count and duration
Exception count by type after deployment
End-to-end transaction drill-down using operation_Id
Trace samples for one failing request path

Application Insights Integration¶

Data Flow Diagram¶

Instrumentation Methods¶

Configuration Examples¶

Enabling Auto-instrumentation via CLI¶

KQL Query Examples¶

Slowest Requests¶

Dependency Failures¶

Application Exceptions¶

Correlate Slow Requests With Dependencies¶

Review Live Failure Trend¶

Recommended Integration Pattern¶

CLI Configuration Workflow¶

1. Create an Application Insights resource¶

2. Retrieve the connection string¶

3. Validate that settings were applied to the app¶

Validation Queries After Enablement¶

Confirm request ingestion¶

Confirm trace ingestion from application logs¶

Confirm exception details include operation correlation¶

Practical Alert Examples¶

Scheduled query alert for dependency failure burst¶

Scheduled query alert for unhandled exceptions¶

What Good Telemetry Looks Like¶

Common Pitfalls¶

Workbook Suggestions¶

See Also¶

Sources¶