Docs

Volcano Engine Function as a Service

Overview

Product features

veFaaS is a fully managed and Serverless computing platform that is driven by events. It provides three product patterns: function, application, and cloud sandbox. This topic describes the core features currently supported by veFaaS by product pattern.

Category	Feature	Description
Application	Application template	An application is a deployment unit for complete business scenarios. It orchestrates functions and dependent resources and supports quick template deployment.
		Managing applications	Creates, deploys, updates, and deletes applications. Functions and resources under an application are managed in a centralized manner.
		Starting or stopping applications	Starts or stops applications as needed and flexibly controls running status and resource usage.
		Automatically building and deploying applications	Automatically build images and deploy applications after code repositories are associated to achieve continuous delivery.
Function	Managing functions	Creates, deploys, updates, and deletes functions. Functions can be deployed through code packages and images.
		Function configuration	Configures the instance memory size, maximum number of instances, concurrency per instance, execution timeout, VPC network, storage, environment variable, and more.
		Trigger	Supports various triggers such as Timer, APIG, TOS, Kafka, RocketMQ, and BMQ to meet the requirements of different event-triggered scenarios.
		Autoscaling	Automatically scales in or out based on request volume and provisions instances to reduce cold-start latency.
		Releasing functions	Releases code and configurations based on version snapshots. Supports full release and canary release. Monitors the release process and manages releases and version records.
		Dashboard overview	Provides metric monitoring, persistent logging (TLS), real-time instance logs, instance WebShell, and online testing to support function observability and debugging.
		Function resource statistics	Collects statistics on resource allocation for accounts and individual functions. Maximum function resources = Instance memory specification × Maximum number of instances. This facilitates appropriate resource allocation and cost control.
		Function type	Event function: Event-triggered functions. They are applicable to event processing, data forwarding, and similar scenarios. Web app function: Functions for hosting web services. They are applicable to web backends and API services. Microservices app function: Functions deployed in the form of microservices. They are applicable to business architectures with service-oriented decomposition. Task function: Responds to invocation events through asynchronous execution, returns a request ID upon task submission, and provides task status tracking and observability capabilities. GPU function: Provides high‑parallelism computing power based on Serverless GPU. This type of function is applicable to AI model inference, training, and media acceleration (audio/video and graphics/image).
Cloud sandboxes	Sandbox application	Creates and manages sandbox applications, which provide runtime templates and configuration for sandbox instances.
		Sandbox instance	Creates, suspends, resumes, and deletes sandbox instances and supports file system snapshots.
		Sandbox image	Manages the images required for sandbox execution and specifies custom sandbox environments.
		Sandbox instance warm pools	Pre‑creates and maintains a certain number of sandbox instances to reduce cold‑start latency.

Last updated: 2026.08.19 14:06:05