docs: update architecture documentation and manifests
Refine technical documentation to reflect current architecture, component structure, and routing conventions for the vcon console.
This commit is contained in:
1 parent
bb938c33c0
commit
367dd819f8
64 files changed
+910
-2661
No files matched your search
@@ -1,12 +1,14 @@
|
||||
# Database Engine Manifest
|
||||
|
||||
## Purpose
|
||||
Database layer managing PostgreSQL schema via Prisma ORM.
|
||||
Manages the shared PostgreSQL connection pool (`pg.Pool`), automatic metadata table initialization, and live database health probes (`core/Database/database.ts`).
|
||||
|
||||
## Absolute Constraint
|
||||
PostgreSQL is utilized ONLY for:
|
||||
1. Client SDK registrations and API keys
|
||||
2. Admin settings and feature flags
|
||||
3. System audit logs
|
||||
## Managed PostgreSQL Tables
|
||||
PostgreSQL is dedicated strictly to platform metadata and security state:
|
||||
1. `admin_users`: Administrator account credentials (scrypt password hashes).
|
||||
2. `admin_login_attempts`: Brute-force login attempt counters and lockout timestamps.
|
||||
3. `client_sdks`: Registered SDK clients, signed JWT tokens, rate limits, status (`active` / `revoked`), and event counters.
|
||||
4. `admin_settings`: Global runtime settings (`max_events_per_sec`, `max_payload_size_kb`, `batch_flush_threshold`, `batch_flush_interval_ms`, `file_rotation_mb`, `jwt_secret`, `enable_idempotency`, `auto_prune_days`).
|
||||
|
||||
**Collected realtime data is NEVER stored per-row in PostgreSQL.** It is stored exclusively in rotated CSV files at `/storage/data/{clientId}/{YYYY-MM-DD}.csv`.
|
||||
## Storage Separation Rule
|
||||
Collected event streams are **never** inserted per-row into PostgreSQL; they are written exclusively in batches to partitioned CSV files under `/storage/data/{clientId}/{YYYY-MM-DD}.csv`.
|
||||
Reference in new issue
Block a user