Skip to content

Latest commit

 

History

History
160 lines (112 loc) · 5.61 KB

FEATURES.md

File metadata and controls

160 lines (112 loc) · 5.61 KB

Extended Features

This document covers the infrastructure and optional integrations that power Iced Latte. Each section explains what the component does, how it is configured, and which free-tier providers work out of the box — so you can run the full stack without spending anything.

Table of Contents

PostgreSQL

The app uses PostgreSQL as its primary database, managed via Liquibase migrations.

Schema

All tables are created and seeded automatically on startup via Liquibase (drop-first: true — the schema is recreated fresh on every start, which is intentional for this open-source project).

Table Purpose
user_details User accounts and profiles
user_granted_authority Roles (ROLE_USER, ROLE_ADMIN)
address User addresses
delivery_address Per-order delivery addresses
product Product catalog
shopping_cart / shopping_cart_item Cart state per user
order / order_item Placed orders
product_review / product_review_likes Reviews and likes
favorite_product / favorite_product_item Favorites list
payment Stripe payment records
file_metadata S3 file references
login_attempts Brute-force tracking
audit_log Audit trail (created/updated timestamps)

Connection

Configured via environment variables (see local.env):

DATASOURCE_HOST=localhost
DATASOURCE_PORT=5432
DATASOURCE_NAME=iced-latte
DATASOURCE_USERNAME=postgres
DATASOURCE_PASSWORD=postgres

Connection pool: HikariCP, min 2 / max 10 connections. SSL required in production (sslmode=require in the JDBC URL).

Seed data

15 users are seeded on every startup. All share the password p@ss1logic11. Test login: olivia@example.com.

Object Storage (AWS S3 / Supabase)

The app uses AWS S3 SDK v2 for file storage. Any S3-compatible provider works — including Supabase Storage, which is free-tier and S3-compatible.

What is stored

Bucket Contents
AWS_PRODUCT_BUCKET Product images (path: ./products)
AWS_USER_BUCKET User avatar images

Presigned URLs are generated per request with a 1-hour expiry and cached in Redis (or Caffeine) for 50 minutes.

How to configure (Supabase)

  1. Create a project at supabase.com
  2. Go to Storage → Buckets and create two buckets: one for products, one for user avatars
  3. Go to Project Settings → API and copy the S3 credentials

Set these environment variables:

AWS_ACCESS_KEY=<supabase-s3-access-key>
AWS_SECRET_KEY=<supabase-s3-secret-key>
AWS_REGION=<your-supabase-region>         # e.g. eu-west-1
AWS_ENDPOINT_URL=https://<project-ref>.supabase.co/storage/v1/s3
AWS_PUBLIC_URL_BASE=https://<project-ref>.supabase.co/storage/v1/object/public
AWS_PRODUCT_BUCKET=<your-products-bucket-name>
AWS_USER_BUCKET=<your-avatars-bucket-name>
AWS_DEFAULT_PRODUCT_IMAGES_PATH=./products

To disable S3 entirely (e.g. for local dev without file uploads):

AWS_ENABLED=false

Redis Cache

Redis is an optional dependency. The app runs fully without it — all caches fall back to in-memory alternatives automatically.

What is cached

Cache Key TTL Fallback
Single product productById::{uuid} 10 min Caffeine
Product brands brands::brandNames 24 h Caffeine
Product sellers sellers::sellerNames 24 h Caffeine
S3 image URL imageUrl::{s3key} 50 min Caffeine
Email verification token email:token:{email} per token TTL Guava
Email send rate email:expiry:{email} per token TTL Guava
JWT blacklist jwt:blacklist:{token} remaining token TTL ConcurrentHashMap
Rate limiter rate:{ip} sliding window Caffeine token bucket

How to enable

Set these environment variables (see local.env for defaults):

REDIS_HOST=your-redis-host
REDIS_PORT=6379
REDIS_PASSWORD=your-password   # leave empty if no auth
REDIS_SSL_ENABLED=false        # set true for TLS (e.g. Upstash)

When REDIS_HOST is not set, the app starts normally with in-memory caches — no configuration needed for local development.

Upstash (free managed Redis)

Upstash offers a free-tier Redis compatible with this setup. Use the TCP endpoint (not the REST URL):

REDIS_HOST=<your-db>.upstash.io
REDIS_PORT=6380
REDIS_PASSWORD=<your-upstash-password>
REDIS_SSL_ENABLED=true

MongoDB (Idea)

💡 This is a future idea, not yet implemented.

All current data in Iced Latte is relational and tightly joined — orders, users, cart, reviews all rely on foreign keys and transactions, which makes PostgreSQL the right fit for them.

The one place MongoDB genuinely fits is the audit log. The existing audit_log PostgreSQL table is append-only, never joined, and has no fixed schema requirement — exactly what a document store is good at. Moving it to MongoDB would mean:

  • No schema migrations when new auditable fields are added
  • Append-only writes with no locking contention on the main DB
  • Free tier on MongoDB Atlas is well-suited for this volume

What would change

  • Add spring-boot-starter-data-mongodb dependency
  • Replace the audit_log PostgreSQL table with a MongoDB collection
  • Use @Document instead of @Entity for the audit entity
  • Configure MONGODB_URI env var pointing to Atlas

Everything else (products, orders, users, cart, reviews) stays in PostgreSQL.