hvt-nearest-atf

A Serverless Node Express lambda (NearestAtfFunction) for showing nearest ATF centres with their availability.

Requirements

• node v18.16.0
• Docker
• SAM CLI

Run Locally

1. Follow build steps in hvt-data to prepare local dataset
2. hvt-geolocation and hvt-read-api must be running
3. npm i
4. cp .env.development .env
5. npm run build:dev
6. npm run start:dev
7. Go to http://localhost:3007/ on browser

Run and watch Locally

As steps above but instead of build:dev

• npm run watch:dev

And in a separate terminal, run

• npm run start:dev

Note: only .ts files are being watched. Any changes to template njk files will not be watched so will not rebuilt.

Debug Locally (VS Code only)

1. Run lambdas in debug mode: npm run start:dev -- -d 5858
2. Add a breakpoint to the lambda being tested (src/controllers/index.controller.ts)
3. Run the debug config from VS Code that corresponds to lambda being tested (NearestAtfFunction)
4. Go to http://localhost:3007/ on browser

Tests

• The Jest framework is used to run tests and collect code coverage
• To run the tests, run the following command within the root directory of the project: npm test
• Coverage results will be displayed on terminal and stored in the coverage directory
  • The coverage requirements can be set in jest.config.js

Build for Production

1. npm i
2. Add environment variables to .env
3. npm run build:prod
4. Zip file and minified assets and can be found in ./dist/

Logging

By using a utility wrapper (src/util/logger) surrounding console.log, the awsRequestId and a "correlation ID" is output with every debug/info/warn/error message.

For this pattern to work, every service/lambda must forward their correlation ID to subsequent services via a header e.g. X-Correlation-Id.

In practice, the first lambda invoked by an initial request will not have received the X-Correlation-Id header, so its correlationId gets defaulted to its lambdaRequestId. This correlationId should then be used when invoking subsequent lambdas via the X-Correlation-Id header. Every lambda called subsequently will then check for that X-Correlation-Id header and inject it into their logs.

This shows an example of what the log looks like from the first invoked lambda:

2020-09-10T17:03:04.891Z	5ff37fce-5ace-114c-9120-a1406cc8d11d	INFO	{"apiRequestId":"c6af9ac6-7b61-11e6-9a41-93e8deadbeef","correlationId":"5ff37fce-5ace-114c-9120-a1406cc8d11d","message":"Here's a gnarly info message from lambda 1 - notice how my correlationId has been set to my lambdaRequestId?"}

This shows an example of what the logs look like from the second invoked lambda (called via the first lambda):

2020-09-10T17:05:31.627Z	32ff455b-057d-1dd7-98b8-7034bf182dc8	INFO	{"apiRequestId":"d9222e0a-6bd9-49e0-84dd-ffe0680bd141","correlationId":"5ff37fce-5ace-114c-9120-a1406cc8d11d","message":"Here's a gnarly info message from lambda 2 - notice how my correlationId is the same as the lambda 1"}