Production changes

Scalability

• App is currently deployed in 1 region and all ECS tasks are running in one AZ, for world wide good performance and HA would deploy it to multiple regions and AZs
• Would have auto-scaling to handle surge in the traffic without human intervention

Observability

• Due to time restrictions added basic metrics which allows me to get RPS and latency. In production environment there will be metric / trace interceptors that would allow us to have more observable system
• Didn't add as it wasn't in spec, but would generate a request ID for each incoming request and refer to that in all logs

Security

• As it wasn't in the spec I didn't implement it, but ideally this API endpoint should be accesible with a per client token, so we could know who are the clients, who are the heavy hitters
• Since it would require me to have a custom domain and a certificate, didn't implement TLS. Would run the service with TLS in production.
• In addition to rate limiter we could put a CDN for DDos protection and better performance

API Gateway

Code

The code for the API Gateway is in api-gateway directory.

Description

Works as a transparent API Gateway between https://polygon-rpc.com. Forwards the requests and returns the response as we get it.

Assumptions

• We're only integrated with only one network. If needed to support multiple networks and make changes to integrate we could have another type of Client in forwarder pkg which implements forwarder.Client interface
• Given I was asked to implement only 2 methods limited the RPC calls that could be done to Polygon to the methods specified
• Given this will be an API exposed to public internet, added a basic rate limiter for all API endpoints

Endpoints

• /rpc: Gets the JSONRPC request as post body, forwards request to polygon and replies with the header + response body. This endpoint has retry mechanism to retry 3 times for the failed requests
• /metrics: Endpoint for Prometheus scrapers to collect metrics for the running process
• /health: Health endpoint for LB usage, it is simplistic at the moment. If we were using a database etc a ping to there could be added just as a simple check.

How to test?

go test ./...

How to run locally?

go run cmd/server/main.go

How to build Docker container?

docker build -t trust-wallet-homework .

Terraform

Assumptions

• Due to time restriction didn't implement TLS connection to ALB as it requires a custom domain and a certificate
• Given this is just deploying one application didn't implement separate modules for VPC, ECS and LB. On a larger Terraform code base abstracting these away with a module would be more maintainable

How to deploy new version of the container?

Push new version

aws ecr get-login-password --region eu-west-1 | docker login --username AWS --password-stdin 471112544726.dkr.ecr.eu-west-1.amazonaws.com
docker build -t api-gateway .
docker tag api-gateway:latest 471112544726.dkr.ecr.eu-west-1.amazonaws.com/api-gateway:latest
docker push 471112544726.dkr.ecr.eu-west-1.amazonaws.com/api-gateway:latest

Trigger deploy of latest version

aws ecs update-service --cluster api-gateway-cluster --service api-gateway-server-service --force-new-deployment

Is this deployed somewhere?

Yes, to my personal AWS account. Can be tested with

curl --header "Content-type: application/json" --request POST --data '{
  "jsonrpc": "2.0",
  "method": "eth_getBlockByNumber",
  "params": [
    "0x134e82a",
    true
  ],
  "id": 2
}'  http://api-gateway-lb-2119393760.eu-west-1.elb.amazonaws.com/rpc