Performance Tests - RAM disk (#17)

Author: Michael Solomon

Diff for: .nvmrc

@@ -0,0 +1 @@
+14.15.0

Diff for: example-application/data-access/order-repository.js

@@ -28,7 +28,6 @@ module.exports = class OrderReposiroty {
   }
 
   async addOrder(orderDetails) {
-    
     return await orderModel.create(orderDetails);
   }
 };

Diff for: example-application/test/docker-compose.yml

@@ -1,4 +1,4 @@
-version: "3"
+version: "3.6"
 services:
   db:
     image: postgres:11
@@ -10,10 +10,4 @@ services:
     container_name: "postgres-for-testing"
     ports:
       - "54320:5432"
-    volumes:
-      - data:/var/lib/postgresql/data
-volumes:
-  data:
-    driver_opts:
-      type: tmpfs
-      device: tmpfs
+    tmpfs: /var/lib/postgresql/data

Diff for: various-receipes/performance/readme.md

@@ -0,0 +1,15 @@
+# Tests Performance
+
+## In-Memory DB
+TL;DR: don't bother
+
+A common time-consuming operation on a typical server is database query, and slow tests are bad.
+Database queries are expensive, also because the data resides on the drive, which is very slowly relative to other parts of the computer.
+One method to speed up the database for testing and local-development is to run the database on RAM, which is way faster than the drive.
+
+But is it worth it? Probably not, and this is why:
+1. SSD is pretty much a standard now, and they are blazingly fast compared to the old HHD.
+2. It's preferred and easier to configure your database for non-durable settings to reduce database-related I/O drastically.
+3. Usually, test scenarios don't involve much data, so the DB's I/O-speed becomes secondary.
+4. The performance between in-memory and regular database with the correct settings is neglectable (if any). 
+Detailed benchmarks and results can be found [here](https://github.com/testjavascript/integration-tests-a-z/issues/9#issuecomment-710674437).