»Invoke Services from Lambda Functions

This topic describes how to invoke services in the mesh from Lambda functions registered with Consul.

Introduction

The following steps describe the process:

1. Deploy the destination service and mesh gateway.
2. Deploy the Lambda extension layer.
3. Deploy the Lambda registrator.
4. Write the Lambda function code.
5. Deploy the Lambda function.
6. Invoke the Lambda function.

You must add the consul-lambda-extension extension as a Lambda layer to enable Lambda functions to send requests to mesh services. Refer to the AWS Lambda documentation for instructions on how to add layers to your Lambda functions.

The layer runs an external Lambda extension that starts a sidecar proxy. The proxy listens on one port for each upstream service and upgrades the outgoing connections to mTLS. It then proxies the requests through to mesh gateways.

Prerequisites

You must deploy the destination services and mesh gateway prior to deploying your Lambda service with the consul-lambda-extension layer.

Deploy the destination service

There are several methods for deploying services to Consul service mesh. The following example configuration deploys a service named static-server with Consul on Kubernetes.

kind: Service
apiVersion: v1
metadata:
  # Specifies the service name in Consul.
  name: static-server
spec:
  selector:
    app: static-server
  ports:
    - protocol: TCP
      port: 80
      targetPort: 8080
---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: static-server
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: static-server
spec:
  replicas: 1
  selector:
    matchLabels:
      app: static-server
  template:
    metadata:
      name: static-server
      labels:
        app: static-server
      annotations:
        'consul.hashicorp.com/connect-inject': 'true'
    spec:
      containers:
        - name: static-server
          image: hashicorp/http-echo:latest
          args:
            - -text="hello world"
            - -listen=:8080
          ports:
            - containerPort: 8080
              name: http
      serviceAccountName: static-server

Deploy the mesh gateway

The mesh gateway must be running and registered to the Lambda function’s Consul datacenter. Refer to the following documentation and tutorials for instructions:

• Mesh Gateways between WAN-Federated Datacenters
• Mesh Gateways between Admin Partitions
• Mesh Gateways between Peered Clusters
• Connect Services Across Datacenters with Mesh Gateways

Deploy the Lambda extension layer

The consul-lambda-extension extension runs during the Init phase of the Lambda function execution. The extension retrieves the data that the Lambda registrator has been configured to store from AWS Parameter Store and creates a lightweight TCP proxy. The proxy creates a local listener for each upstream defined in the CONSUL_SERVICE_UPSTREAMS environment variable.

The extension periodically retrieves the data from the AWS Parameter Store so that the function can process requests. When the Lambda function receives a shutdown event, the extension also stops.

1. Download the consul-lambda-extension extension from releases.hashicorp.com:

   curl -o consul-lambda-extension_<version>_linux_amd64.zip https://releases.hashicorp.com/consul-lambda/<version>/consul-lambda-extension_<version>_linux_amd64.zip
   

   Copy

2. Create the AWS Lambda layer in the same AWS region as the Lambda function. You can create the layer manually using the AWS CLI or AWS Console, but we recommend using Terraform:

   consul-lambda-extension.tf

   Copy

   resource "aws_lambda_layer_version" "consul_lambda_extension" {
       layer_name       = "consul-lambda-extension"
       filename         = "consul-lambda-extension_<version>_linux_amd64.zip"
       source_code_hash = filebase64sha256("consul-lambda-extension_<version>_linux_amd64.zip")
       description      = "Consul service mesh extension for AWS Lambda"
   }
   

Deploy the Lambda registrator

Configure and deploy the Lambda registrator. Refer to the registrator configuration documentation and the registrator deployment documentation for instructions.

Write the Lambda function code

Refer to the AWS Lambda documentation for instructions on how to write a Lambda function. In the following example, the function calls an upstream service on port 2345:

package main

import (
    "context"
    "io"
    "fmt"
    "net/http"
    "github.com/aws/aws-lambda-go/lambda"
)    

type Response struct {
    StatusCode int  `json:"statusCode"`
    Body    string `json:"body"`
}

func HandleRequest(ctx context.Context, _ interface{}) (Response, error) {
    resp, err := http.Get("http://localhost:2345")
    fmt.Println("Got response", resp)
    if err != nil {
     return Response{StatusCode: 500, Body: "Something bad happened"}, err
    }

    if resp.StatusCode != 200 {
     return Response{StatusCode: resp.StatusCode, Body: resp.Status}, err
    }

    defer resp.Body.Close()

    b, err := io.ReadAll(resp.Body)
    if err != nil {
     return Response{StatusCode: 500, Body: "Error decoding body"}, err
    }

    return Response{StatusCode: 200, Body: string(b)}, nil
}

func main() {
    lambda.Start(HandleRequest)
}

Deploy the Lambda function

1. Create and apply an IAM policy that allows the Lambda function’s role to fetch the Lambda extension’s data from the AWS Parameter Store. The following example, creates an IAM role for the Lambda function, creates an IAM policy with the necessary permissions and attaches the policy to the role:

   lambda-iam-policy.tf

   Copy

   resource "aws_iam_role" "lambda" {
   name = "lambda-role"
   
   assume_role_policy = <<EOF
   {
   "Version": "2012-10-17",
   "Statement": [
       {
       "Action": "sts:AssumeRole",
       "Principal": {
           "Service": "lambda.amazonaws.com"
       },
       "Effect": "Allow",
       "Sid": ""
       }
   ]
   }
   EOF
   }
   
   resource "aws_iam_policy" "lambda" {
   name        = "lambda-policy"
   path        = "/"
   description = "IAM policy lambda"
   
   policy = <<EOF
   {
   "Version": "2012-10-17",
   "Statement": [
       {
       "Effect": "Allow",
       "Action": [
           "ssm:GetParameter"
       ],
       "Resource": "arn:aws:ssm:*:*:parameter${local.this_lambdas_extension_data_path}"
       }
   ]
   }
   EOF
   }
   
   resource "aws_iam_role_policy_attachment" "lambda" {
   role    = aws_iam_role.lambda.name
   policy_arn = aws_iam_policy.lambda.arn
   }
   

2. Configure and deploy the Lambda function. Refer to the Lambda extension configuration reference for information about all available options. There are several methods for deploying Lambda functions. The following example uses Terraform to deploy a function that can invoke the static-server upstream service using mTLS data stored under the /lambda_extension_data prefix:

   lambda-function.tf

   Copy

   resource "aws_lambda_function" "example" {
   …
   function_name = "lambda"
   role = aws_iam_role.lambda.arn
   tags = {
       "serverless.consul.hashicorp.com/v1alpha1/lambda/enabled" = "true"
   }
   variables = {
       environment = {
         CONSUL_MESH_GATEWAY_URI = var.mesh_gateway_http_addr
         CONSUL_SERVICE_UPSTREAMS = "static-server:2345:dc1"
         CONSUL_EXTENSION_DATA_PREFIX = "/lambda_extension_data"
       }
   }
   layers = [aws_lambda_layer_version.consul_lambda_extension.arn]  
   

3. Run the terraform apply command and Consul automatically configures a service for the Lambda function.

Lambda extension configuration

Define the following environment variables in your Lambda functions to configure the Lambda extension. The variables apply to each Lambda function in your environment:

Invoke the Lambda function

If intentions are enabled in the Consul service mesh, you must create an intention that allows the Lambda function's Consul service to invoke all upstream services prior to invoking the Lambda function. Refer to Service Mesh Intentions for additional information.

There are several ways to invoke Lambda functions. In the following example, the aws lambda invoke CLI command invokes the function:

$ aws lambda invoke --function-name lambda /dev/stdout | cat