lambda-http for AWS Lambda in Rust

lambda-http is an abstraction that takes HTTP-shaped payloads from different AWS services and turns them into standard http objects, making it easy to write Lambda functions in Rust that serve HTTP traffic regardless of the upstream trigger.

lambda-http handler is made of:

• Request - Represents an HTTP request
• IntoResponse - Future that will convert an [IntoResponse] into an actual [LambdaResponse]

We are able to handle requests from:

• API Gateway REST, HTTP and WebSockets API lambda integrations
• AWS ALB
• AWS Lambda function URLs
• AWS VPC Lattice

Thanks to the Request type we can seamlessly handle all of these triggers without having to write service-specific code.

There is also an extension for lambda_http::Request structs that provides access to API Gateway, ALB, and VPC Lattice features.

For example some handy extensions:

• query_string_parameters - Return pre-parsed http query string parameters, parameters provided after the ? portion of a url associated with the request
• path_parameters - Return pre-extracted path parameters, parameter provided in url placeholders /foo/{bar}/baz/{qux} associated with the request
• lambda_context - Return the Lambda context for the invocation; see the runtime docs
• request_context - Return the ALB, API Gateway, or VPC Lattice request context
• payload - Return the Result of a payload parsed into a type that implements serde::Deserialize

See the lambda_http::RequestPayloadExt and lambda_http::RequestExt traits for more extensions.

Examples

Here you will find a few examples to handle basic scenarios:

• Reading a JSON from a body and deserialize into a structure
• Reading query string parameters
• Lambda Request Authorizer
• Passing the Lambda execution context initialization to the handler

Reading a JSON from a body and deserialize into a structure

The code below creates a simple API Gateway proxy (HTTP, REST) that accepts in input a JSON payload.

use lambda_http::{run, http::{StatusCode, Response}, service_fn, Error, IntoResponse, Request, RequestPayloadExt};
use serde::{Deserialize, Serialize};
use serde_json::json;

#[tokio::main]
async fn main() -> Result<(), Error> {
    tracing_subscriber::fmt()
        .without_time()
        .with_max_level(tracing::Level::INFO)
        .init();

    run(service_fn(function_handler)).await
}

pub async fn function_handler(event: Request) -> Result<impl IntoResponse, Error> {
    let body = event.payload::<MyPayload>()?;

    let response = Response::builder()
        .status(StatusCode::OK)
        .header("Content-Type", "application/json")
        .body(json!({
            "message": "Hello World",
            "payload": body,
          }).to_string())
        .map_err(Box::new)?;

    Ok(response)
}

#[derive(Deserialize, Serialize, Debug, Clone)]
pub struct MyPayload {
    pub prop1: String,
    pub prop2: String,
}

Reading query string parameters

use lambda_http::{run, http::{StatusCode, Response}, service_fn, Error, RequestExt, IntoResponse, Request};
use serde_json::json;

#[tokio::main]
async fn main() -> Result<(), Error> {
    tracing_subscriber::fmt()
        .without_time()
        .with_max_level(tracing::Level::INFO)
        .init();

    run(service_fn(function_handler)).await
}

pub async fn function_handler(event: Request) -> Result<impl IntoResponse, Error> {
    let name = event.query_string_parameters_ref()
        .and_then(|params| params.first("name"))
        .unwrap_or_else(|| "stranger")
        .to_string();

    // Represents an HTTP response
    let response = Response::builder()
        .status(StatusCode::OK)
        .header("Content-Type", "application/json")
        .body(json!({
            "message": format!("Hello, {}!", name),
          }).to_string())
        .map_err(Box::new)?;

    Ok(response)
}

Lambda Request Authorizer

Because lambda-http is an abstraction, we cannot use it for the Lambda Request Authorizer case. If you remove the abstraction, you need to handle the request/response for your service.

use aws_lambda_events::apigw::{
    ApiGatewayCustomAuthorizerRequestTypeRequest, ApiGatewayCustomAuthorizerResponse, ApiGatewayCustomAuthorizerPolicy, IamPolicyStatement,
};
use lambda_runtime::{run, service_fn, Error, LambdaEvent};
use serde_json::json;

#[tokio::main]
async fn main() -> Result<(), Error> {
    tracing_subscriber::fmt()
        .without_time()
        .with_max_level(tracing::Level::INFO)
        .init();

    run(service_fn(function_handler)).await
}

pub async fn function_handler(event: LambdaEvent<ApiGatewayCustomAuthorizerRequestTypeRequest>) -> Result<ApiGatewayCustomAuthorizerResponse, Error> {
    // do something with the event payload
    let method_arn = event.payload.method_arn.unwrap();
    // for example we could use the authorization header
    if let Some(token) = event.payload.headers.get("authorization") {
        // do something

        return Ok(custom_authorizer_response(
            "ALLOW",
            "some_principal",
            &method_arn,
        ));
    }

    Ok(custom_authorizer_response(
      &"DENY".to_string(),
      "",
      &method_arn))
}

pub fn custom_authorizer_response(effect: &str, principal: &str, method_arn: &str) -> ApiGatewayCustomAuthorizerResponse {
    let stmt = IamPolicyStatement {
        action: vec!["execute-api:Invoke".to_string()],
        resource: vec![method_arn.to_owned()],
        effect: Some(effect.to_owned()),
    };
    let policy = ApiGatewayCustomAuthorizerPolicy {
        version: Some("2012-10-17".to_string()),
        statement: vec![stmt],
    };
    ApiGatewayCustomAuthorizerResponse {
        principal_id: Some(principal.to_owned()),
        policy_document: policy,
        context: json!({ "email": principal }), // https://github.com/aws/aws-lambda-rust-runtime/discussions/548
        usage_identifier_key: None,
    }
}

Passing the Lambda execution context initialization to the handler

One of the best practices is to take advantage of execution environment reuse to improve the performance of your function. Initialize SDK clients and database connections outside the function handler. Subsequent invocations processed by the same instance of your function can reuse these resources. This saves cost by reducing function run time.

use aws_sdk_dynamodb::model::AttributeValue;
use chrono::Utc;
use lambda_http::{run, http::{StatusCode, Response}, service_fn, Error, RequestExt, IntoResponse, Request};
use serde_json::json;

#[tokio::main]
async fn main() -> Result<(), Error> {
    tracing_subscriber::fmt()
        .without_time()
        .with_max_level(tracing::Level::INFO)
        .init();

    let config = aws_config::from_env()
        .load()
        .await;

    let dynamodb_client = aws_sdk_dynamodb::Client::new(&config);

    run(service_fn(|event: Request| function_handler(&dynamodb_client, event))).await
}

pub async fn function_handler(dynamodb_client: &aws_sdk_dynamodb::Client, event: Request) -> Result<impl IntoResponse, Error> {
    let table = std::env::var("TABLE_NAME").expect("TABLE_NAME must be set");

    let name = event.query_string_parameters_ref()
        .and_then(|params| params.first("name"))
        .unwrap_or_else(|| "stranger")
        .to_string();

    dynamodb_client
        .put_item()
        .table_name(table)
        .item("ID", AttributeValue::S(Utc::now().timestamp().to_string()))
        .item("name", AttributeValue::S(name.to_owned()))
        .send()
        .await?;

    // Represents an HTTP response
    let response = Response::builder()
        .status(StatusCode::OK)
        .header("Content-Type", "application/json")
        .body(json!({
            "message": format!("Hello, {}!", name),
          }).to_string())
        .map_err(Box::new)?;

    Ok(response)
}

Integration with API Gateway stages

When you integrate HTTP Lambda functions with API Gateway stages, the path received in the request will include the stage as the first segment, for example /production/api/v1, where production is the API Gateway stage.

If you don't want to receive the stage as part of the path, you can set the environment variable AWS_LAMBDA_HTTP_IGNORE_STAGE_IN_PATH to true, either in your Lambda function configuration, or inside the main Rust function. Following the previous example, when this environment variable is present, the path that the function receives is /api/v1, eliminating the stage from the first segment.

Feature flags

lambda_http is a wrapper for HTTP events coming from four different AWS services: Application Load Balancer (ALB), API Gateway, Lambda Function URLs, and VPC Lattice. API Gateway can send events from three different endpoint types: REST APIs, HTTP APIs, and WebSockets. lambda_http transforms events from all these sources into native http::Request objects, so you can incorporate Rust HTTP semantics into your Lambda functions.

By default, lambda_http compiles your function to support any of those services. This increases the compile time of your function because we have to generate code for all the sources. In reality, you'll usually put a Lambda function only behind one of those sources. You can choose which source to generate code for with feature flags.

The available feature flags for lambda_http are the following:

• alb: for events coming from Amazon Elastic Load Balancer.
• apigw_rest: for events coming from Amazon API Gateway REST APIs.
• apigw_http: for events coming from Amazon API Gateway HTTP APIs and AWS Lambda Function URLs.
• apigw_websockets: for events coming from Amazon API Gateway WebSockets.
• vpc_lattice: for events coming from AWS VPC Lattice.
• catch-all-fields: enables catch-all fields in event types to preserve unknown JSON fields during deserialization.
• builders: enables builder pattern support for request/response types in aws_lambda_events, useful for constructing test fixtures.

If you only want to support one of these sources, you can disable the default features, and enable only the source that you care about in your package's Cargo.toml file. Substitute the dependency line for lambda_http for the snippet below, changing the feature that you want to enable:

[dependencies.lambda_http]
version = "0.5.3"
default-features = false
features = ["apigw_rest"]