Microsoft Docs

Azure Functions Host and Scale

This article explains the Azure Functions host and how it scales to meet demand.

The Azure Functions Host

The Azure Functions host is the process that runs your function code. It is responsible for:

• Detecting and loading function code.
• Handling trigger events and invoking functions.
• Managing the execution context of functions.
• Communicating with other Azure services.
• Providing a local development experience.

When you deploy your Azure Functions application, the host is automatically provisioned and managed by Azure. You don't need to worry about managing servers or operating systems.

Scaling Behavior

Azure Functions automatically scales the number of function app instances based on incoming event traffic. This automatic scaling is a key benefit of the serverless model.

Consumption Plan

The Consumption plan offers automatic scaling and pay-per-execution pricing. When your application receives events, Azure automatically scales out the number of instances to handle the load. When the load decreases, instances are scaled in. This means you only pay for the compute time you consume.

Premium Plan

The Premium plan provides pre-warmed instances to reduce cold start latency and offers VNet connectivity. It also offers more predictable scaling behavior compared to the Consumption plan for certain workloads.

App Service Plan

With an App Service plan, your functions run on dedicated virtual machines. You have more control over the scaling behavior and can configure auto-scaling rules based on metrics like CPU or memory usage. However, you are billed for the underlying VMs regardless of execution, and scaling is not as instantaneous as in the Consumption or Premium plans.

Scaling Factors

The scaling of your Azure Functions application is primarily influenced by:

• Incoming event rate: The volume and frequency of events that trigger your functions.
• Execution duration: How long each function execution takes. Longer executions can hold up instances.
• Concurrency: The number of events being processed simultaneously.
• Instance limits: There are default limits on the number of instances per function app, which can be adjusted.

Monitoring Scaling

You can monitor the scaling behavior of your function app using Azure Monitor and Application Insights. Key metrics to watch include:

• Function Execution Count: Tracks the number of times your functions are invoked.
• Function Execution Units (Consumption plan): Measures the compute time consumed.
• Average Memory Working Set: Indicates memory usage per instance.
• HTTP Server Errors: Can indicate issues with scaling or resource constraints.

Host Configuration

The Azure Functions host can be configured using the host.json file. This file allows you to control aspects like:

• Logging levels and output.
• Extension configuration.
• HTTP settings.
• Queue processing.

For example, you can set the maximum number of concurrent executions for certain triggers:

{
  "version": "2.0",
  "extensions": {
    "queues": {
      "maxDequeueCount": 5,
      "newBatchSize": 10,
      "vmLatencyHint": 3000
    }
  },
  "logging": {
    "applicationInsights": {
      "samplingSettings": {
        "isEnabled": true,
        "excludedTypes": "Request"
      }
    }
  }
}
            

For more detailed information, refer to the official Azure Functions documentation.