MSA Design and Implementation - 01
I designed the MSA project that I will be working on
System Configuration
I drew a system configuration diagram for what I'm aiming for.
k8s Configuration
It consists of a total of 3 nodes, with 2 workers where the actual apps run. The app and api configurations are clear, but the DB server configuration was unclear because I was wondering if it's appropriate to run a DB server that needs to be stateful on k8s, which is specialized for providing stateless services.
Therefore, I plan to set taints so that it doesn't receive k8s failure recovery or scheduling, and use the replica functionality provided by MongoDB for backup and HA configuration.
Configuration Based on Ingress
Users are ingress-ed to /app, /api paths by nginx-ingress, and app and api will each have 2 replicas created for HA configuration. At this time, the unique point is that when the WAS expressed as 'app' performs DB queries or business logic processing, it always requests to the api server. This configuration enables API provision to 3rd party apps and will help create an extensible configuration where even if a new-app is created later, it won't depend on the existing old-app.
What to Do Next
- I will implement app and api servers with express while studying k8s in parallel.
- Let's clearly understand the concept and implementation of mongo replication.
appendix. If I get the chance, I want to apply front-end libraries like React while developing the app.