Skip to main content

FaaS Use Cases

Overview​

Function as a Service (FaaS) enables developers to deploy and execute code without managing servers or infrastructure. Its event-driven and scalable architecture makes it suitable for a wide variety of workloads, from lightweight APIs to large-scale AI inference services.

API Services​

Build and deploy REST APIs, internal services, and backend endpoints without managing application servers.

Common examples: user management APIs, authentication services, billing APIs, inventory services, internal platform APIs.

Recommended runtimes: Python 3.x [Http], Python 3.11 [FastAPI], Node.js [Http]

Benefits:

  • Automatic scaling based on traffic
  • Independent deployment per service
  • Rapid development and iteration
  • No infrastructure overhead

Webhooks​

Receive and process events from external platforms.

Common examples: payment gateway callbacks, Git repository events, CRM integrations, messaging platform events, CI/CD notifications.

Recommended runtimes: Python 3.x [Http], Node.js [Http]

Benefits:

  • Fast deployment
  • Event-driven execution — only runs when triggered
  • Cost-efficient processing

Backend for Frontend (BFF)​

Create lightweight backend services that aggregate data for web and mobile applications.

Common examples: mobile application backends, dashboard APIs, data aggregation services, GraphQL gateways.

Recommended runtimes: Python 3.11 [FastAPI], Node.js [Http]

Microservices​

Deploy independent services that can be developed, scaled, and maintained separately.

Common examples: user service, notification service, order service, payment service, analytics service.

Recommended runtimes: Python 3.11 [FastAPI], Node.js [Http], Go 1.21 [Http]

Benefits:

  • Independent scaling per service
  • Fault isolation — one service failure does not cascade
  • Faster, independent deployments
  • Technology flexibility per service

AI and Machine Learning Inference​

Serve trained machine learning models through HTTP APIs.

Common examples: text generation, image classification, object detection, recommendation engines, NLP applications.

Recommended runtimes: Python PyTorch 2.1.2 [FastAPI], Python TensorFlow 2.19.0 [FastAPI], Custom Container Image

Hardware: Use GPU functions for accelerated inference. GPU runtimes support PyTorch and TensorFlow, or bring your own container.

Data Processing​

Execute transformation and processing workloads on demand.

Common examples: CSV processing, ETL pipelines, data enrichment, data validation, report generation.

Recommended runtimes: Python 3.x [Http], Python 3.11 [FastAPI]

Scheduled Tasks​

Run recurring background jobs and maintenance workflows.

Common examples: cleanup jobs, report generation, data synchronization, cache refresh operations, usage aggregation.

Recommended runtimes: Python 3.x [Http], Node.js [Http]

AI Agents and Automation​

Deploy autonomous workflows and agent-based applications.

Common examples: support assistants, workflow orchestration, document processing, knowledge retrieval, automated operations.

Recommended runtimes: Python 3.11 [FastAPI], Python PyTorch 2.1.2 [FastAPI], Custom Container Image

Custom Runtime Applications​

Deploy existing containerized workloads with minimal modification.

Common examples: legacy applications, custom frameworks, specialized runtimes, internal tooling.

Recommended runtime: Custom Container Image

Choosing the Right Runtime​

ScenarioRecommended Runtime
New REST APIPython 3.11 [FastAPI] or Node.js 24 [Http]
Existing Express appNode.js 20 [Express]
Lightweight webhook or functionPython 3.x [Http] or Node.js [Http]
AI / ML inference (GPU)Python PyTorch [FastAPI] or Python TensorFlow [FastAPI]
Full custom applicationCustom Container Image
.NET workloadC# / .NET 7.0 [Http]
PHP applicationPHP 8.2 [Http]

Image Processing with E2E Object Storage​

This example shows how to combine E2E FaaS with E2E Object Storage to process images stored in a bucket.

Scenario​

A business needs to process user-provided images to enhance their quality before storage or publication — tasks such as adjusting contrast, resizing, or applying specific transformations.

Steps​

Step 1: Access Object Storage

  • Connect to the specified object storage bucket using the provided access credentials.
  • Retrieve the image object from the designated bucket.

Step 2: Enhance the Image

  • Read the content of the retrieved image object.
  • Open the image using the Pillow library.
  • Apply the desired enhancement (e.g., increase contrast) using the ImageEnhance module.
  • Convert the enhanced image back to binary data for uploading.

Why FaaS works well here:

  • Image processing runs asynchronously without blocking the user experience.
  • Resources are consumed only when processing is triggered — no idle servers.
  • The function scales automatically to handle bursts of image upload events.
Last updated on June 3, 2026.