Database Schema & Relational Indexing Strategy

You are a Principal Database Architect and PostgreSQL internals expert.

Design or review the database schema and indexing strategy for this feature:

Business Requirements & Query Workflows:
```
Queries:
1. Fetch recent activity for a tenant ordered by timestamp:
   SELECT * FROM audit_logs WHERE tenant_id = ? AND action = ? ORDER BY created_at DESC LIMIT 50;
2. Search audit logs by JSON metadata fields.
3. Purge logs older than 90 days efficiently.
```

Proposed Schema DDL:
```sql
CREATE TABLE audit_logs (
  id SERIAL PRIMARY KEY,
  tenant_id VARCHAR(50),
  action VARCHAR(100),
  metadata JSONB,
  created_at TIMESTAMP
);
```

Target Scale: 25M rows/month, 90% writes, retention period 180 days
RDBMS: PostgreSQL 16

Provide an architectural schema & indexing review:
1. **Schema Normalization & Data Types**:
   - Check appropriate data types (e.g. `TIMESTAMPTZ` vs `TIMESTAMP`, `UUIDv7` vs `SERIAL`, `NUMERIC` for currency).
   - Flag denormalization risks or missing foreign key constraints.
2. **Indexing Architecture (B-Tree, GIN, BRIN, Partial)**:
   - Identify index requirements for high-frequency `WHERE`, `ORDER BY`, and `JOIN` clauses.
   - Design composite indexes respecting the leftmost prefix rule (Equality first, Range second).
   - Suggest partial indexes (`WHERE status != 'ARCHIVED'`) to save disk and memory.
3. **Partitioning & Sharding Readiness**:
   - Recommend table partitioning strategies (by range/date or hash) if table scale warrants it.
4. **Complete DDL Implementation**:
   - Provide the complete, production-ready SQL migration script with constraints and indexes.