MySQL 慢查询日报(线上 prod8591)2026-09-08

统计概览

项目值
统计范围2026-09-08 00:00:00 ~ 23:59:59(Asia/Shanghai)
环境prod-8591
慢日志条数(ES mysql-slow_*)78,463 条(指纹聚合取样 23,116 条,覆盖率 29%,超过 2 万条上限截断)
Top10 单条最慢4,719.91 s(user=bi,ZHDX_t8591_107)
非备份模式最大累计3.98h(402 次 × 35.62 s)
备份模式(user 含 backup/databasus_backup)65 条、累计 1.55h(单独标注,不触发深析)
数据来源ES mysql-slow_*(@timestamp)+ MCP 只读 MySQL(describe_schema/explain)
生成时间2026-09-09 11:32(Asia/Shanghai)

Top10 明细(按单条耗时)

#耗时(s)库.表行数 sent/examineduser / server慢因分类
14,719.91None.t8_mall_stat161,049,194 / 322,098,388bi / ZHDX_t8591_107其他(需结合 SQL 与执行计划人工判断)
23,796.81None.t8_mall_stat161,219,059 / 322,438,118bi / ZHDX_t8591_107其他(需结合 SQL 与执行计划人工判断)
32,953.58None.t8_mall_stat161,286,024 / 322,572,048bi / ZHDX_t8591_107其他(需结合 SQL 与执行计划人工判断)
42,923.39(未解析)65 / 32,017,786bi / ZHDX_t8591_107扫描行数远大于返回行数(32,017,786/65)
52,878.72None.t8_mall_stat161,152,675 / 322,305,350bi / ZHDX_t8591_107其他(需结合 SQL 与执行计划人工判断)
62,711.62(未解析)2,918,535 / 36,460,687bi / ZHDX_t8591_107扫描行数远大于返回行数(36,460,687/2,918,535)
72,388.73(未解析)1,136,551 / 9,674,136bi / ZHDX_t8591_107其他(需结合 SQL 与执行计划人工判断)
82,258.27(未解析)1,135,937 / 9,665,276bi / ZHDX_t8591_107其他(需结合 SQL 与执行计划人工判断)
92,176.44(未解析)64 / 32,017,785bi / ZHDX_t8591_107扫描行数远大于返回行数(32,017,785/64)
101,574.68(未解析)845 / 6,767,873bi / ZHDX_t8591_107扫描行数远大于返回行数(6,767,873/845)

高频慢查询模式(指纹聚合,取样 23,116 条)

按条数 Top5

#模式(指纹摘要)条数平均(s)累计serveruser
1SELECT c.* FROM t8_user_id_card_number_list AS c WHERE c.img_type = ? AND c.status = ? AND3,2353.503.14hZHDX_t8591_143,ZHDX_t8591_16,ZHDX_t8591_a8591
2select session, from, body from chat_messages where from = ? and time > ? and 2,7521.311.00hZHDX_t8591_15im8591
3select t1_design_img.id, t1_design_img.views, t1_design_img.collection, t1_de`2,5721.3357.0minZHDX_t8591_73a591
4select count(distinct session) as num from chat_messages where from = ? and time > ?2,1471.3347.5minZHDX_t8591_15im8591
5SELECT id FROM ware_item WHERE id > ? AND status > ? AND (actionTime BETWEEN ? AND ? )1,0801.4225.5minZHDX_t8591_61a8591

按累计耗时 Top5

#模式条数平均(s)累计serveruser
1select sum(searched_times) as total, keyword from t8_keyword_statistics where date bet40235.623.98hZHDX_t8591_15a8591
2SELECT c.* FROM t8_user_id_card_number_list AS c WHERE c.img_type = ? AND c.status = ? AND3,2353.503.14hZHDX_t8591_143,ZHDX_t8591_16,ZHDX_t8591_a8591
3select wallet, sum(if(type = ?, amount, ?)) - sum(if(type = ?, amount, ?)) as balance from15350.522.15hZHDX_t8591_61a8591
4SELECT * FROM t8_mall_stat ORDER BY created_at DESC;23799.322.11hZHDX_t8591_107bi
5SELECT a.buyUser AS buyerId, a.sellUser AS sellerId, c.order_number AS orderNo, a.id AS tx22217.361.23hZHDX_t8591_107bi
备份模式全天 65 条 / 1.55h,未触发深度分析(如条数环比异常翻倍再人工介入)。

触发深度分析的模式(条数>100 或 累计>30min,非备份,最多 8 个)

1. select sum(searched_times) as total, keyword from t8_keyword_statistics where date between ? and —— 402 次 × 35.62s = 3.98h

2. SELECT c.* FROM t8_user_id_card_number_list AS c WHERE c.img_type = ? AND c.status = ? AND c.image ! —— 3,235 次 × 3.50s = 3.14h

3. select wallet, sum(if(type = ?, amount, ?)) - sum(if(type = ?, amount, ?)) as balance from a8591.` —— 153 次 × 50.52s = 2.15h

4. SELECT * FROM t8_mall_stat ORDER BY created_at DESC; —— 2 次 × 3799.32s = 2.11h

5. SELECT a.buyUser AS buyerId, a.sellUser AS sellerId, c.order_number AS orderNo, a.id AS txNo, c.orde —— 2 次 × 2217.36s = 1.23h

6. select session, from, body from chat_messages where from = ? and time > ? and type = ? —— 2,752 次 × 1.31s = 1.00h

7. select t1_design_img.id, t1_design_img.views, t1_design_img.collection, t1_design_img. —— 2,572 次 × 1.33s = 57.0min

8. SELECT id, user_id, money, ware_id, type, ware_type_limit, amount_limit, client_limit, active_id, co —— 2 次 × 1538.55s = 51.3min

逐条优化建议汇总

优先级对象建议
P1?.t8_keyword_statistics(402×35.6s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P2?.t8_user_id_card_number_list(3,235×3.5s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P3a8591.t8_money_trade_log(153×50.5s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P4?.t8_mall_stat(2×3799.3s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P5?.ware_item(2×2217.4s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P6?.chat_messages(2,752×1.3s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P7?.t1_design_img(2,572×1.3s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
P8?.t8_user_voucher(2×1538.6s)结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
—Top1 ?.t8_mall_stat结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
—Top2 ?.t8_mall_stat结合 SQL 文本与执行计划人工复核; 考虑降频/缓存
—Top3 ?.t8_mall_stat结合 SQL 文本与执行计划人工复核; 考虑降频/缓存

*采集脚本 .agents/scripts/mysql_slow_prod_report.py,中间数据 outputs/slow_report_prod8591_20260908.json;全程只读。*