Skip to main content

External Lambda Functions

When running Medplum in AWS, Medplum uses AWS Lambdas to execute bot logic.

By default, Medplum manages the full AWS Lambda life cycle of create, update, deploy, and execute using a predefined AWS Lambda naming convention.

In a self hosted environment, you may want use your own AWS Lambda life cycle, your own naming conventions, or a third party management tool such as Serverless Framework or SST. Medplum calls these "external functions".

This document describes how to enable and use external functions.

Enable​

Enable external functions by setting the botCustomFunctionsEnabled config setting to true.

For example, if using AWS Parameter Store for your configuration settings, you would create a new parameter called /medplum/{environment}/botCustomFunctionsEnabled with value true.

For more details, see the Config Settings guide.

Configure the bot​

After the botCustomFunctionsEnabled config setting is set to true, the server will look for a FHIR Identifier with system https://medplum.com/bot-external-function-id for the AWS Lambda function name.

For example, consider this Bot:

{
  "resourceType": "Bot",
  "identifier": [
    {
      "system": "https://medplum.com/bot-external-function-id",
      "value": "my-custom-lambda-function"
    }
  ],
  "name": "External Demo Bot",
  "runtimeVersion": "awslambda"
}

Deploy the bot​

When using external functions, the Medplum server is no longer involved with deploying the AWS Lambda. The $deploy endpoint is effectively a no-op.

Execute the bot​

Executing the bot is done exactly the same as with a normal Medplum-managed AWS Lambda, using either the $execute endpoint or a FHIR Subscription.