MyBatis 性能优化最佳实践：从 SQL 到连接池的全面调优指南

🚀 MyBatis 性能优化最佳实践：从 SQL 到连接池的全面调优指南

⚡ 一、性能优化全景图

💡 MyBatis 性能瓶颈分析

优化目标​​：

• 🚀 降低数据库查询耗时
• 📉 减少网络IO和磁盘IO
• 💾 优化内存使用效率
• 🔄 提高并发处理能力

🗃️ 二、SQL 与查询优化

💡 SQL 优化核心策略

1. 索引优化实践
   ​​场景​​：用户表按状态和创建时间查询

-- 优化前：全表扫描
SELECT * FROM users WHERE status = 1 ORDER BY create_time DESC;-- 优化后：添加复合索引
CREATE INDEX idx_status_createtime ON users(status, create_time DESC);-- 优化SQL：利用索引覆盖
SELECT id, name, email FROM users 
WHERE status = 1 
ORDER BY create_time DESC 
LIMIT 100;

​​索引优化建议​​：

场景	索引策略	效果
等值查询	单列索引	快速定位
范围查询	范围列放在复合索引后面	避免索引失效
排序操作	排序字段加索引	避免filesort
多条件查询	复合索引	索引覆盖

2. 解决 N+1 查询问题
   ​​问题场景​​

// N+1查询示例：1次查询用户 + N次查询订单
public List<User> getUsersWithOrders() {List<User> users = userMapper.selectAllUsers();for (User user : users) {// 每次循环都执行一次查询List<Order> orders = orderMapper.selectByUserId(user.getId());user.setOrders(orders);}return users;
}

​​解决方案​​：

<!-- 使用连接查询一次性获取 -->
<select id="selectUsersWithOrders" resultMap="userWithOrdersMap">SELECT u.*, o.id as order_id, o.amount, o.create_time as order_timeFROM users uLEFT JOIN orders o ON u.id = o.user_idWHERE u.status = 1
</select><resultMap id="userWithOrdersMap" type="User"><id property="id" column="id"/><result property="name" column="name"/><collection property="orders" ofType="Order"><id property="id" column="order_id"/><result property="amount" column="amount"/><result property="createTime" column="order_time"/></collection>
</resultMap>

3. 懒加载配置优化

<!-- mybatis-config.xml -->
<settings><!-- 开启懒加载 --><setting name="lazyLoadingEnabled" value="true"/><setting name="aggressiveLazyLoading" value="false"/><setting name="lazyLoadTriggerMethods" value=""/>
</settings><!-- 特定关联的懒加载配置 -->
<resultMap id="userLazyMap" type="User"><collection property="orders" ofType="Order" select="selectOrdersByUserId" column="id"fetchType="lazy"/>
</resultMap>

📊 SQL 优化效果对比

优化策略	优化前耗时	优化后耗时	提升幅度
索引优化	1200ms	150ms	8倍
N+1解决	N+1次查询	1次查询	N倍
懒加载	立即加载所有数据	按需加载	2-5倍

📦 三、批量操作优化

💡 批量插入性能对比

1. BatchExecutor 批量操作

// 使用BatchExecutor执行批量操作
public void batchInsertUsers(List<User> users) {SqlSession sqlSession = sqlSessionFactory.openSession(ExecutorType.BATCH);try {UserMapper mapper = sqlSession.getMapper(UserMapper.class);for (int i = 0; i < users.size(); i++) {mapper.insertUser(users.get(i));// 每1000条提交一次，避免内存溢出if (i % 1000 == 0 && i > 0) {sqlSession.commit();sqlSession.clearCache(); // 清空缓存避免OOM}}sqlSession.commit();} finally {sqlSession.close();}
}// 批量更新示例
public void batchUpdateUsers(List<User> users) {try (SqlSession sqlSession = sqlSessionFactory.openSession(ExecutorType.BATCH)) {UserMapper mapper = sqlSession.getMapper(UserMapper.class);for (User user : users) {mapper.updateUser(user);}sqlSession.commit();}
}

2. foreach 批量插入

<!-- 批量插入SQL -->
<insert id="batchInsertUsers">INSERT INTO users (name, email, status) VALUES<foreach item="user" collection="list" separator=",">(#{user.name}, #{user.email}, #{user.status})</foreach>
</insert><!-- 批量更新SQL -->
<update id="batchUpdateUsers"><foreach item="user" collection="list" separator=";">UPDATE users SET name = #{user.name}, email = #{user.email}WHERE id = #{user.id}</foreach>
</update>

// 服务层调用
public void batchInsertUsers(List<User> users) {// 分批处理，避免SQL过长int batchSize = 1000;for (int i = 0; i < users.size(); i += batchSize) {List<User> batch = users.subList(i, Math.min(i + batchSize, users.size()));userMapper.batchInsertUsers(batch);}
}

📊 批量操作性能数据

操作方式	1000条数据耗时	内存占用	适用场景
单条插入	15s	低	实时单条插入
BatchExecutor	1.5s	中	中等批量数据
foreach批量	0.8s	高	大数据量导入

🏊 四、连接池深度调优

💡 HikariCP 优化配置

# application.yml
spring:datasource:hikari:# 连接池大小maximum-pool-size: 20minimum-idle: 5# 连接超时与生命周期connection-timeout: 30000idle-timeout: 600000max-lifetime: 1800000# 性能优化参数connection-init-sql: SELECT 1connection-test-query: SELECT 1validation-timeout: 3000# 监控相关register-mbeans: trueleak-detection-threshold: 60000

🔧 连接池参数调优指南

参数	建议值	说明	影响
maximum-pool-size	CPU核心数 * 2 + 1	最大连接数	并发能力
minimum-idle	maximum-pool-size / 2	最小空闲连接	响应速度
connection-timeout	30000ms	连接获取超时	系统韧性
idle-timeout	600000ms	空闲连接超时	资源回收
max-lifetime	1800000ms	连接最大生命周期	连接 freshness

📈 连接池监控配置

// 连接池监控端点
@Bean
public MeterRegistryCustomizer<MeterRegistry> metricsCustomizer(DataSource dataSource) {return registry -> {if (dataSource instanceof HikariDataSource) {HikariDataSource hikariDataSource = (HikariDataSource) dataSource;// 注册监控指标new HikariDataSourceMetrics(hikariDataSource, "app-datasource").bindTo(registry);}};
}// 健康检查配置
management:endpoints:web:exposure:include: health,metrics,infoendpoint:health:show-details: alwaysprobes:enabled: true

💾 五、缓存策略优化

💡 多级缓存架构

1. 二级缓存优化配置

<!-- Mapper缓存配置 -->
<cacheeviction="LRU"flushInterval="60000"size="1024"readOnly="true"blocking="false"/><!-- 特定语句缓存配置 -->
<select id="selectUserById" resultType="User" useCache="true" flushCache="false">SELECT * FROM users WHERE id = #{id}
</select>

2. Redis 分布式缓存集成

// Redis缓存配置
@Configuration
@EnableCaching
public class RedisConfig extends CachingConfigurerSupport {@Beanpublic RedisTemplate<String, Object> redisTemplate(RedisConnectionFactory factory) {RedisTemplate<String, Object> template = new RedisTemplate<>();template.setConnectionFactory(factory);// 使用Jackson序列化Jackson2JsonRedisSerializer<Object> serializer = new Jackson2JsonRedisSerializer<>(Object.class);ObjectMapper mapper = new ObjectMapper();mapper.setVisibility(PropertyAccessor.ALL, JsonAutoDetect.Visibility.ANY);mapper.activateDefaultTyping(mapper.getPolymorphicTypeValidator(), ObjectMapper.DefaultTyping.NON_FINAL);serializer.setObjectMapper(mapper);template.setValueSerializer(serializer);template.setKeySerializer(new StringRedisSerializer());return template;}@Beanpublic CacheManager cacheManager(RedisConnectionFactory factory) {RedisCacheConfiguration config = RedisCacheConfiguration.defaultCacheConfig().entryTtl(Duration.ofHours(1)) // 缓存1小时.disableCachingNullValues() // 不缓存null值.serializeValuesWith(RedisSerializationContext.SerializationPair.fromSerializer(new GenericJackson2JsonRedisSerializer()));return RedisCacheManager.builder(factory).cacheDefaults(config).build();}
}

3. 缓存使用示例

// 服务层缓存注解
@Service
public class UserService {@Cacheable(value = "users", key = "#id")public User getUserById(Long id) {return userMapper.selectById(id);}@CacheEvict(value = "users", key = "#user.id")public void updateUser(User user) {userMapper.updateUser(user);}@Caching(evict = {@CacheEvict(value = "users", key = "#user.id"),@CacheEvict(value = "user-list", allEntries = true)})public void updateUserWithCache(User user) {userMapper.updateUser(user);}
}

📊 缓存性能对比

缓存级别	平均响应时间	适用场景	注意事项
一级缓存	0.1ms	会话内重复查询	数据实时性要求高
二级缓存	0.5ms	跨会话共享数据	需要处理缓存一致性
Redis缓存	1.2ms	分布式环境	网络开销需要考虑

🎯 六、综合实践与监控

💡 全链路性能监控

// SQL执行监控切面
@Aspect
@Component
@Slf4j
public class SqlPerformanceAspect {private final ThreadLocal<Long> startTime = new ThreadLocal<>();@Around("execution(* com.example.mapper.*.*(..))")public Object monitorSqlPerformance(ProceedingJoinPoint joinPoint) throws Throwable {startTime.set(System.currentTimeMillis());try {return joinPoint.proceed();} finally {long cost = System.currentTimeMillis() - startTime.get();String methodName = joinPoint.getSignature().getName();// 记录慢查询if (cost > 1000) {log.warn("Slow SQL detected: {} - {}ms", methodName, cost);// 发送到监控系统Metrics.counter("sql.slow.query").increment();}// 记录性能指标Metrics.timer("sql.execute.time").record(cost, TimeUnit.MILLISECONDS);startTime.remove();}}
}

🛡️ 生产环境配置模板

# 生产环境MyBatis优化配置
mybatis:configuration:# 性能相关配置cache-enabled: truelazy-loading-enabled: trueaggressive-lazy-loading: falsemultiple-result-sets-enabled: trueuse-column-label: trueuse-generated-keys: true# 执行器配置default-executor-type: REUSEdefault-statement-timeout: 30# 其他优化map-underscore-to-camel-case: truelocal-cache-scope: SESSION# 数据源配置
spring:datasource:hikari:maximum-pool-size: 20minimum-idle: 5connection-timeout: 30000idle-timeout: 600000max-lifetime: 1800000leak-detection-threshold: 60000

📈 性能监控指标体系

监控指标	预警阈值	处理策略
SQL执行时间	> 1000ms	优化SQL或添加索引
连接池等待时间	> 500ms	调整连接池大小
缓存命中率	< 80%	优化缓存策略
批量操作耗时	> 预期2倍	调整批量大小或策略

🔚 总结与延伸

📚 优化要点回顾

1. SQL优化​​：索引、避免N+1、合理使用连接查询
2. ​​批量操作​​：BatchExecutor、foreach批量插入
3. 连接池​​：HikariCP参数调优、监控配置
4. 缓存策略​​：多级缓存、分布式缓存集成

🚀 持续优化建议

​​优化循环​​：监控 → 分析 → 优化 → 验证 → 持续改进