blead / lazy-loading-reg

AWS Lambda function handler implementation for querying structured course data from Chulalongkorn University's Office of the Registrar website, with caching.

Lazy Loading Reg

Notes about VPC

It is possible to put Lambda instances into a VPC to access private services such as an ElastiCache instance. Though this will create more overhead for Lambda and also require additional setups to access the internet through a NAT gateway.

Please refer to these official guides for more details:

• Tutorial: Configuring a Lambda Function to Access Amazon ElastiCache in an Amazon VPC
• How can I grant internet access to my VPC Lambda function?

Alternatively, if the structure does not require a private network, the Lambda functions can just use no VPC.

Project setup

Required: Node.js.

1. Install dependencies.

  npm install

2. Zip the directory to be deployed with AWS Lambda.

zip -r ../lazy-loading-reg.zip ./*

Create Lambda functions

1. Create 2 new lambda functions (for 2 endpoints at /courses and /courses/:course-number).

• Function names can be anything. For clarity, courseList and courseInformation are preferred and used in this example.
• Node.js 8.10 runtime is preferred.

2. Upload the zip file to both functions. Set the following Handler names:

• For courseList: courseList.handler.
• For courseInformation: courseInformation.handler.

3. Set environment variables:

Name	Default Value	Required	Description
CACHING	false	no	If not set, caching will be disabled. Set to memory to use in-memory caching, and redis to use Redis backend.
CACHE_TTL	259200000 (3 days)	no	Cache TTL in milliseconds. Used for both in-memory cache and Redis.
CACHE_SIZE		no	Set a maximum number of entries for the cache before pruning expired entries. Only used for in-memory cache.
REDIS_URL		yes	The URL of the Redis server. Format: [redis[s]:]//[[user][:password@]][host][:port][/db-number][?db=db-number[&password=bar[&option=value]]]. See Node Redis documentation and Redis URI Schemes at IANA for more details. Only used for Redis.

4. Set execution role.

• The role should have access to Lambda.
• If using VPC, the role should also have VPC access.

5. Modify basic settings.

• From testing, 256MB of allocated memory should be enough to function.
• Set a moderately high timeout to account for Reg's response time, e.g., 1-2 minutes.

Create API Gateway

1. Create a regional REST API by importing swagger.json.

2. Set integration for each endpoint.

• At /courses/GET, set integration to the courseList Lambda function. Use Lambda Proxy integration.
• At /courses/{courseNumber}/GET, set integration to the courseInformation Lambda function. Use Lambda Proxy integration.

3. For each resource, click Actions > Enable CORS > Enable CORS and replace existing CORS headers.

4. Deploy the API. Click Actions > Deploy API. Set the desired stage name.

Configure Lambda functions

1. Add API Gateway as trigger for each Lambda function. Set security to open.

2. Permissions might need to be modified to properly allow access to /courses endpoint. (The default permissions allow /courseList endpoint instead. Genius.)

• Click on Permissions.
• Permissions will be shown in JSON. To properly execute the function, AWS:SourceArn needs to be set to arn:aws:execute-api:region:id:api-id/*/GET/courses. If the default permissions do not match, a new permission can be created with AWS CLI.

  aws lambda add-permission \
  --function-name courseList \
  --statement-id some-id \
  --action lambda:InvokeFunction \
  --principal apigateway.amazonaws.com \
  --source-arn arn:aws:execute-api:region:id:api-id/*/GET/courses

• The same can be done with /courses/:course-number but the default permissions set by AWS probably work fine.
• Refer to the official tutorial for more information.

Redis

A simple Redis instance can be used as a caching backend. If Redis is open to public, always set a password to secure the data.

To use ElastiCache or a private Redis instance, see notes about VPC at the beginning of this document.