POSTGRESQL CLIENT IMPLEMENTATION IN AWS LAMBDA CONTAINER IMAGES

NOTE: This is a GPT-4o optimized version of the blog. For a more human-friendly version, see

TLDR: Use amazon-linux-extras to install PostgreSQL 10 client in Lambda container images for Postgres database operations.

CORE COMPONENTS

Base Image:

- Image: public.ecr.aws/lambda/provided:al2

- OS: Amazon Linux 2

- Package Manager: yum + amazon-linux-extras

Required Packages:

- amazon-linux-extras

- postgresql10

- jq (optional, for JSON processing)

IMPLEMENTATION STEPS

1. Dockerfile Configuration:

FROM public.ecr.aws/lambda/provided:al2

RUN yum update -y && \

yum install -y amazon-linux-extras && \

amazon-linux-extras install postgresql10 && \

yum install -y jq

2. SAM Template Configuration:

Resources:

DBSetupFunction:

Type: AWS::Serverless::Function

Properties:

PackageType: Image

ImageConfig:

Command: ["/var/runtime/bootstrap"]

Environment:

Variables:

DB_HOST: !GetAtt AuroraCluster.Endpoint.Address

DB_PORT: !GetAtt AuroraCluster.Endpoint.Port

IMPLEMENTATION DETAILS

Environment Configuration:

1. Database Credentials

- DB_HOST

- DB_PORT

- DB_USER

- DB_PASSWORD

Network Configuration:

1. VPC Requirements

VpcConfig:

SecurityGroupIds:

- !Ref LambdaSecurityGroup

SubnetIds: !Ref SubnetIds

Security Configuration:

1. Security Group Rules

SecurityGroupIngress:

- IpProtocol: tcp

FromPort: 5432

ToPort: 5432

ERROR RESOLUTION MATRIX

Error | Cause | Solution

----------------------|------------------|------------------

Missing redhat-release| Wrong repo | Use amazon-linux-extras

Connection timeout | VPC config | Check security groups

Command not found | PATH issues | Verify postgresql10 installation

Runtime conflict | PackageType:Image| Remove Runtime field, use ImageConfig instead

DEPLOYMENT PROCESS

1. Build Command:

sam build

2. Deployment Command:

sam deploy --guided

3. Verification Command:

aws lambda invoke \

--function-name ${STACK_NAME}-DBSetupFunction-${RANDOM_SUFFIX} \

--payload '{"seed": true}' \

--log-type Tail \

--query 'LogResult' \

--output text response.json | base64 -d

TECHNICAL REQUIREMENTS

Minimum Configuration:

1. Memory: 512MB

2. Timeout: 300 seconds

3. PackageType: Image (Note: Cannot be used with Runtime field)

Network Requirements:

1. VPC access

2. Internet access for package installation

TESTING PROTOCOL

Local Testing:

1. Command: sam local invoke DBSetupFunction

Production Testing:

1. Command:

aws lambda invoke \

--function-name ${STACK_NAME}-DBSetupFunction-${RANDOM_SUFFIX} \

--payload '{"seed": true}' \

--log-type Tail \

--query 'LogResult' \

--output text response.json | base64 -d

Note: Replace ${STACK_NAME} with your SAM stack name and ${RANDOM_SUFFIX} with the suffix AWS generates during deployment.

You can find the full function name in the AWS Console or using:

aws lambda list-functions --query 'Functions[?starts_with(FunctionName, `your-stack-name`)].FunctionName'

REFERENCE IMPLEMENTATION

Complete Directory Structure:

.

├── Dockerfile.migrations

├── template.yaml

├── lambda/

│ └── bootstrap

└── migrations/

└── seed.sql

Required Files:

1. Dockerfile.migrations

2. template.yaml

3. bootstrap script

4. migration scripts

MINIMAL BOOTSTRAP EXAMPLE

/lambda/bootstrap:

```

#!/bin/bash

# Initialize error handling

set -euo pipefail

# Handler function

function handler () {

EVENT_DATA=$1

# Set default values if not provided

DB_USER=${DB_USER:-postgres}

DB_PASSWORD=${DB_PASSWORD:-postgres}

DB_HOST=${DB_HOST:-localhost}

DB_PORT=${DB_PORT:-5432}

# Run simple query

RESPONSE=$(PGPASSWORD="${DB_PASSWORD}" psql -h "${DB_HOST}" -p "${DB_PORT}" -U "${DB_USER}" -d postgres -t -c "SELECT NOW();")

echo $RESPONSE

}

# Main loop

while true

do

# Get an event

HEADERS="$(mktemp)"

EVENT_DATA=$(curl -sS -LD "$HEADERS" -X GET "http://${AWS_LAMBDA_RUNTIME_API}/2018-06-01/runtime/invocation/next")

REQUEST_ID=$(grep -Fi Lambda-Runtime-Aws-Request-Id "$HEADERS" | tr -d '[:space:]' | cut -d: -f2)

# Execute the handler function

RESPONSE=$(handler "$EVENT_DATA")

# Send the response

curl -X POST "http://${AWS_LAMBDA_RUNTIME_API}/2018-06-01/runtime/invocation/$REQUEST_ID/response" -d "$RESPONSE"

done

```

Note: Make script executable with: chmod +x lambda/bootstrap