An experimental and contrived preview of Golang. A very simple event driven API to fetch information about the Author, Michael Pearce.
A 'Services' approach - Separate lambda functions for each category of information contact, summary, interests etc. Including all of the basic actions - get, put, delete etc.
curl "https://api.michaelpearce.info/contact?channel=Email"
curl "https://api.michaelpearce.info/contact?channel=Facebook"
curl "https://api.michaelpearce.info/contact?channel=Instagram"
curl "https://api.michaelpearce.info/contact?channel=LinkedIn"
curl "https://api.michaelpearce.info/contact?channel=Website"
curl "https://api.michaelpearce.info/summary?section=Interests"
curl "https://api.michaelpearce.info/summary?section=Certifications"
curl "https://api.michaelpearce.info/summary?section=Examples"
curl "https://api.michaelpearce.info/summary?section=Code"
Build on services provided by Amazon Web Services (AWS) cloud.
- api gateway, for a scalable backend
- lambda, to run 'serverless' code functions in response to events from API Gateway
- dynamodb, for data store
- cloudfront, for CDN
- route 53 for domain name registration and hosting, and CNAME record set
- cloudwatch for logging and cloudwatch/sns for alarms
The project is build on AWS Infrastructure and assumes the infrastructure mentioned in the above 'Infrastructure' section is already deployed and configured to run the code. Along with the correct permissions!
go get github.com/aws/aws-lambda-go/lambda
go get github.com/aws/aws-sdk-go
go get github.com/aws/aws-lambda-go/events
Set Go OS and Architecture variables so that it is built to run on the Linux servers that AWS Lambda uses. Build it generally wherever you want, but the temp folder keeps things tidy in the long run. the name of the binary file should represent the main function name that is initially called to trigger the function, in this case main. The final input parameter is the location of the go project that you want to build, assuming it can load the package from your $GOPATH.
env GOOS=linux GOARCH=amd64 go build -o /tmp/main contact
To use environment variables at run time I've added a function getEnv in the main file that takes a key and a fallback. You can call this function with the key you are expecting and if that environment variable is not set, use the fallback. This can be used for things like Stage, Region, etc.
To use with Lambda, the environment variables will need to be set on creation or by updating the lambda function to include the variable and it's value. Similarly if you are running the code locally you would need to set the variable with a value, or the fallback (passed into the function in the code) is used.
Package it up as a zip file for Lambda.
zip -j /tmp/main.zip /tmp/main
Deploy to Lambda, however you prefer. For a quick proof of concept you can use the CLI.
aws lambda update-function-code --function-name contact --zip-file fileb:///tmp/main.zip
If the aws command times out because the package is too big (and your network bandwith is low), it is easier to copy to S3 first, then deploy from there.
aws s3 cp /tmp/main.zip s3://<YOUR-S3-BUCKET-NAME>/
aws lambda update-function-code --function-name contact --s3-bucket <YOUR-S3-BUCKET-NAME> --s3-key main.zip