支持一次性任务与按消耗计费时，何时优先选用Azure Functions而非ACI？

Great question! Since both services handle one-off tasks and use consumption-based billing, the choice often comes down to your specific workflow, development needs, and integration requirements. Here are the key scenarios where Azure Functions is the better pick:

• Built-in event-driven triggers are non-negotiable
  Azure Functions comes with a huge library of native triggers (Blob Storage, Event Hubs, HTTP, Timer, Service Bus, etc.) that let you kick off tasks automatically without extra setup. For example, if you need to process a file as soon as it’s uploaded to Blob Storage, you can configure a Blob trigger in minutes—no need to build a separate triggering system (like Logic Apps or a custom scheduler) that would be required for ACI.

• Your task is lightweight and code-focused
  If you’re running small, discrete code snippets (think: a Python script to transform data, a JavaScript function to send a notification), Functions eliminates the overhead of building, maintaining, and pushing Docker images. You can write code directly in the Azure portal, VS Code, or your IDE of choice and deploy in seconds—no container configuration needed. ACI shines for complex workloads with custom dependencies, but for simple code tasks, Functions is far more streamlined.

• Automatic, hassle-free scaling is critical
  Functions natively scales out based on incoming load—adds more instances as demand spikes and scales down to zero when idle. With ACI, you’d have to set up custom autoscale rules via Azure Monitor or pair it with Azure Container Apps to get similar behavior, which adds extra complexity. For bursty, unpredictable workloads (like processing a flood of incoming messages), Functions’ auto-scaling is a huge win.

• You need tight integration with Azure’s serverless ecosystem
  Functions plays seamlessly with other Azure serverless tools: use Durable Functions for stateful workflow orchestration, pair it with Logic Apps for low-code workflow building, or connect directly to Azure Cosmos DB, Azure SQL, or Azure Storage with pre-built bindings. ACI can integrate with these services too, but you’ll have to write more custom code to handle connections and workflow coordination.

• Cost efficiency for short, frequent tasks
  While both use consumption billing, Functions’ pricing is granular (billed per execution, down to the millisecond) and includes a generous free tier. For tasks that run in seconds and fire frequently, this can be more cost-effective than ACI, where you’re billed for container runtime (even if the task finishes quickly) and may incur longer cold start times for infrequent tasks.

内容的提问来源于stack exchange，提问作者Alberto