Add architecture decision records for Discord.js, PostgreSQL, polling strategy, and modular monolith

This commit is contained in:
2026-05-17 15:51:02 +02:00
parent 77e43a6a38
commit c348204772
5 changed files with 28 additions and 16 deletions
+4 -16
View File
@@ -2,28 +2,16 @@
## ADR-001: Discord.js over Discord.py
**Status:** Accepted
**Context:** Node.js and Python both viable. Team has stronger Node.js experience.
**Decision:** Use Discord.js for the bot runtime.
**Consequences:** Faster development velocity; ecosystem rich with community plugins (command handling, modals, components).
[ADR-001: Discord.js over Discord.py](./09_architecture_decisions/ADR-001.md)
## ADR-002: PostgreSQL over MongoDB
**Status:** Accepted
**Context:** Mileage data is relational (user → role → event → score). MongoDB adds flexibility but sacrifices query power.
**Decision:** Use PostgreSQL with JSONB columns for extensible user metadata.
**Consequences:** Schema migrations needed for new features; joins straightforward.
[ADR-002: PostgreSQL over MongoDB](./09_architecture_decisions/ADR-002.md)
## ADR-003: Polling over Webhooks for Content Sources
**Status:** Accepted
**Context:** YouTube, Instagram, TikTok do not provide push webhooks to arbitrary third parties without approval.
**Decision:** Poll public APIs on a CRON schedule (15-min interval).
**Consequences:** ~5 min delay between publish and Discord post; API quota management required.
[ADR-003: Polling over Webhooks for Content Sources](./09_architecture_decisions/ADR-003.md)
## ADR-004: Modular Monolith over Microservices
**Status:** Accepted
**Context:** Single-developer or small team; deployment complexity of microservices not justified.
**Decision:** Single process with internal module boundaries (commands, events, mileage, syndication).
**Consequences:** Simpler deploy, debug, and develop; can extract services later if needed.
[ADR-004: Modular Monolith over Microservices](./09_architecture_decisions/ADR-004.md)
@@ -0,0 +1,6 @@
# ADR-001: Discord.js over Discord.py
**Status:** Accepted
**Context:** Node.js and Python both viable. Team has stronger Node.js experience.
**Decision:** Use Discord.js for the bot runtime.
**Consequences:** Faster development velocity; ecosystem rich with community plugins (command handling, modals, components).
@@ -0,0 +1,6 @@
# ADR-002: PostgreSQL over MongoDB
**Status:** Accepted
**Context:** Mileage data is relational (user → role → event → score). MongoDB adds flexibility but sacrifices query power.
**Decision:** Use PostgreSQL with JSONB columns for extensible user metadata.
**Consequences:** Schema migrations needed for new features; joins straightforward.
@@ -0,0 +1,6 @@
# ADR-003: Polling over Webhooks for Content Sources
**Status:** Accepted
**Context:** YouTube, Instagram, TikTok do not provide push webhooks to arbitrary third parties without approval.
**Decision:** Poll public APIs on a CRON schedule (15-min interval).
**Consequences:** ~5 min delay between publish and Discord post; API quota management required.
@@ -0,0 +1,6 @@
# ADR-004: Modular Monolith over Microservices
**Status:** Accepted
**Context:** Single-developer or small team; deployment complexity of microservices not justified.
**Decision:** Single process with internal module boundaries (commands, events, mileage, syndication).
**Consequences:** Simpler deploy, debug, and develop; can extract services later if needed.