Database Admin Agent
Optimize queries, design schemas, fix bottlenecks. Your database performance expert for PostgreSQL, MySQL, MongoDB, and more.
When to Use
- Dashboard queries timing out (>5s response)
- Tables growing but queries slowing down
- Need schema design for new features
- Connection pool exhaustion or deadlocks
Key Capabilities
| Area | What It Does |
|---|---|
| Query Optimization | EXPLAIN ANALYZE, index recommendations, rewrite slow queries (45s → 0.8s typical) |
| Schema Design | Normalized tables, optimal indexes, constraints, triggers, partitioning strategies |
| Performance Analysis | Health checks, connection pool tuning, cache hit ratio, bloat detection |
| Backup & HA | WAL archiving, point-in-time recovery, replication setup, failover config |
| Database Systems | PostgreSQL, MySQL/MariaDB, MongoDB, Redis, SQLite, distributed databases |
Common Use Cases
1. Fix Slow Dashboard Query
Who: Backend dev with 45-second query Prompt:
"Dashboard query takes 45 seconds. Optimize it:
SELECT u.name, COUNT(o.id), SUM(o.total)
FROM users u LEFT JOIN orders o ON u.id = o.user_id
WHERE o.created_at >= '2024-01-01'
GROUP BY u.id ORDER BY SUM(o.total) DESC LIMIT 100;"
Output: EXPLAIN ANALYZE, missing index report, optimized query (0.8s), materialized view option, implementation plan
2. Schema Design for New Feature
Who: Product team adding order history
Prompt: "Design schema for order history: users, products, orders, payments. Handle 10M+ orders."
Output: Complete DDL with indexes, constraints, triggers, denormalization strategy, partitioning plan
3. Database Health Check
Who: DevOps investigating production slowness
Prompt: "Run comprehensive health check on production DB"
Output: Connection pool analysis, missing FK indexes, bloat report, slow query top 5, cost analysis, action plan
4. Backup Strategy
Who: SRE setting up disaster recovery
Prompt: "Setup backup strategy with PITR for production PostgreSQL"
Output: Daily full backup script, WAL archiving config, restore procedures, S3 upload, verification tests
5. Connection Pool Exhaustion
Who: Frontend seeing “too many connections” errors
Prompt: "95/100 connections used, app timing out. Fix connection pool."
Output: PgBouncer config, app connection pooling code, max_connections tuning, monitoring queries
Pro Tips
For Query Optimization:
- Always provide table sizes + existing indexes for better analysis
- Agent creates covering indexes (include all SELECT columns)
- Expects
plans/reports/folder for optimization reports
For Schema Design:
- Mention scale requirements (10M rows? 1B?). Affects partitioning strategy
- Agent denormalizes strategically (e.g., product name in order_items for history)
- Foreign key indexes created automatically (common oversight)
For Health Checks:
- Run during business hours (not 3am). Real traffic reveals issues
- Agent checks: connections, slow queries, missing indexes, bloat, cache hit ratio
- Provides immediate + short-term + long-term action plans
Performance Benchmarks:
- Query optimization: Typically 90-98% faster execution time
- Index creation: 15-20 min per 1M rows (concurrent, no downtime)
- Health check: Identifies $4K/month in savings (typical)
Related Agents
- Debugger - Debug application database issues
- Fullstack Developer - Implement schema changes in app
- Tester - Validate database migrations
Key Takeaway
Database Admin agent turns 45-second queries into 0.8-second queries with index strategies and query rewrites. Designs production-ready schemas with proper normalization, constraints, and partitioning. Delivers actionable health reports with specific SQL fixes and cost impact ($4K/month savings typical).