playground

flowchart LR
   load-test --> app-go
   load-test -.-> app-java
   load-test -.-> proxy
   proxy -.-> app-java
   proxy -.-> app-go
   proxy -.-> app-go-1
   proxy -.-> app-go-2

   app-java -- storage --> postgres
   app-java -- worker queues --> rabbitmq
   app-go -- storage --> postgres
   app-go -- worker queues --> rabbitmq
   app-go -- cache --> redis
   app-go -. alternative worker queues .-> broker

note:

1. dotted line refers to future update
2. solid line refers to what we currently have

directory briefs

• ./.devcontainer, contain dev-container definition. reference to vscode's dev-container
• ./.docker, contain definitions/configuration to setup container in docker
• ./.vscode, contain repo's workspace setting for vscode
• ./app-go, credit managements app implemented using golang
• ./app-java, credit managements app implemented using java + spring boot
• ./broker, an attempt to create message-broker in golang
• ./load-test, contain scripts to execute load test scenarios

learning points

1. learn how to make credits management doesnt go kaboom when theres huge transaction intake
   1. this covered inside ./app-go and already fully implemented
   2. additionally I wanted to created ./app-java just to see how it performed compared to golang one. It should contain the same spec as ./app-go
      1. implemented, but only for server, no worker
2. learn how to utilize docker for service orchestration
3. learn how to load test using grafana/k6, which we'll be using to run the scenario + replicate huge intake process
   1. this got scratched because as per writing this, we cannot meddle with grafana/k6 directly with golang
4. build a simple load test library that has observabilities element built in it
5. visualize the observability metrics from the experiment, using:
   1. grafana, to visualize
   2. cadvisor, to scrap resource metrics from the containers
   3. prometheus, for app-go metrics
   4. open-telemetry, identifies bottle-necks and traces
6. utilize rabbitmq for message broker. I'm using this for asynchronous process.

setup

• use vscode, reopen repo in container. Ref: VSCode's dev-container
  • This help isolate the development
• to setup the whole thing, run the following command at ./ once inside dev-container

  make start

links

• grafana dashboard (still in work)