Monitoring Azure Functions

Introduction to Monitoring

Effective monitoring is crucial for understanding the health, performance, and usage of your Azure Functions. It allows you to detect issues early, optimize performance, and ensure your applications are running reliably.

Azure Functions integrates deeply with Azure Monitor, providing comprehensive insights into your function executions, resource utilization, and potential errors.

Key Monitoring Tools and Metrics

Azure Application Insights

Application Insights is the primary tool for monitoring Azure Functions. It provides:

• Live Metrics: Real-time performance data as your functions run.
• Performance Analysis: Detailed metrics on request rates, response times, failure rates, and dependency performance.
• Error Tracking: Automatic detection and reporting of exceptions and failures.
• Usage Analytics: Insights into how users interact with your functions.
• Log Stream: Access to logs generated by your functions.

You can enable Application Insights by configuring it in your Function App's settings or during creation.

Azure Monitor Logs (Log Analytics)

When your Function App is configured with Application Insights, logs are sent to a Log Analytics workspace. This allows you to query your logs using Kusto Query Language (KQL) for deeper analysis and custom troubleshooting.

Common log tables include:

• requests: Information about function invocations.
• exceptions: Details of exceptions thrown by your functions.
• traces: Custom logs generated by your function code (e.g., using context.log()).

-- Example KQL query to find failed function executions in the last hour
requests
| where timestamp > ago(1h)
| where success == false
| project timestamp, name, resultCode, duration
| order by timestamp desc

Azure Monitor Metrics

Azure Monitor provides a rich set of metrics for your Function App, including:

• Function Execution Count
• Function Execution Units (FUs)
• Memory Working Set
• HTTP 2xx/3xx/4xx/5xx Response Codes
• Average Response Time

These metrics can be viewed in the Azure portal and used to set up alerts.

Configuring Logging

Azure Functions can be configured to log different levels of detail:

• Information: Standard execution logs.
• Warning: Potential issues.
• Error: Unrecoverable problems.
• Debug: Detailed information useful for debugging (often enabled selectively).

You can configure the logging level in your Function App's host.json file:

{
  "version": "2.0",
  "logging": {
    "applicationInsights": {
      "samplingSettings": {
        "isEnabled": true,
        "excludedTypes": "Request"
      }
    },
    "logLevel": {
      "default": "Information",
      "Host.Results": "Information",
      "Function.MyFunctionName": "Debug"
    }
  }
}

Using context.log() or ILogger in your function code allows you to emit custom messages that will appear in Application Insights or Log Analytics.

Setting Up Alerts

Proactive alerting is essential for responding quickly to issues. You can configure alerts in Azure Monitor based on metrics and log queries.

Alerting on Metrics

Create alerts for conditions like:

• High number of HTTP 5xx errors.
• High function execution duration.
• Low function execution count (indicating a potential outage).

Alerting on Logs

Create alerts based on KQL queries, for example:

• Alert when a specific error message appears in the logs.
• Alert when the number of failed requests exceeds a threshold within a time window.

Alerts can trigger notifications via email, SMS, webhooks, or other actions.

Best Practices for Monitoring

• Enable Application Insights: Always enable Application Insights for your production Function Apps.
• Use Structured Logging: Log relevant data in a structured format to make queries easier.
• Set Appropriate Log Levels: Use "Information" for most production scenarios and enable "Debug" selectively for troubleshooting.
• Implement Health Checks: Create specific HTTP-triggered functions that act as health endpoints, reporting their status.
• Monitor Dependencies: Track performance of external services your functions interact with.
• Regularly Review Dashboards: Familiarize yourself with your function app's performance and identify trends.
• Set Up Meaningful Alerts: Configure alerts for critical conditions and tune them to avoid alert fatigue.