Performance Schema是MySQL重要的性能監控工具，可以幫助我們找到MySQL的性能瓶頸以及一些其他的SLOW SQL等問題，能有效的幫助我們進行系統調優。下面從以下幾個點對Performance schema以及sys56工具進行說明
- 獲取數據庫執行最多的SQL語句
- 單條執行時間最長的SQL
- 最頻繁的表
- 從未被使用過的索引
- 文件IO消耗
Performance Schema功能簡介
MySQL通過一個單獨的performance schema存儲引擎來實現MySQL運行過程中的性能點監控，文檔中將這些性監控點命名爲“instrucment”。通過
show variables like “performance_schema
可以查看是否開啓了performance schema存儲引擎。
performance schema雖然提供的性能數控豐富，但是不便於人類讀。因此偉大的percona爲我們提供了sys56工具，sys56包含了一堆的view、function以及存儲過程方便我們使用
SYS56安裝
下載 git clone https://github.com/hitYangfei/mysql-sys.git
導入sys56.sql即可
也可以去Percona的官網去下載最新版本，這是我的一個clone
數據庫執行最多的SQL
mysql> select * from sys.statement_analysis limit 1\G
*************************** 1. row ***************************
query: COMMIT
db: zabbix
full_scan:
exec_count: 2954549
err_count: 1
warn_count: 0
total_latency: 35.64h
max_latency: 29.45 s
avg_latency: 43.43 ms
lock_latency: 00:01:17.61
rows_sent: 0
rows_sent_avg: 0
rows_examined: 0
rows_examined_avg: 0
tmp_tables: 0
tmp_disk_tables: 0
rows_sorted: 0
sort_merge_passes: 0
digest: e51be358a1cbf99c1acab35cc1c6b683
first_seen: 2015-02-10 13:30:57
last_seen: 2015-03-10 15:41:55
1 row in set (0.01 sec)通過查詢statement_analysis視圖即可。
數據庫執行時間最長的SQL
mysql> select * from statements_with_runtimes_in_95th_percentile limit 2\G
*************************** 1. row ***************************
query: INSERT INTO t1 VALUES (...)
db: test
full_scan:
exec_count: 3
err_count: 2
warn_count: 0
total_latency: 00:04:11.35
max_latency: 00:02:32.12
avg_latency: 00:01:23.78
rows_sent: 0
rows_sent_avg: 0
rows_examined: 0
rows_examined_avg: 0
first_seen: 2015-03-08 11:32:43
last_seen: 2015-03-08 11:45:15
digest: a0583512c4eb718088979fe23a35a893
*************************** 2. row ***************************
query: SELECT SQL_NO_CACHE * FROM `history`
db: zabbix
full_scan: *
exec_count: 1
err_count: 0
warn_count: 0
total_latency: 40.94 s
max_latency: 40.94 s
avg_latency: 40.94 s
rows_sent: 26657012
rows_sent_avg: 26657012
rows_examined: 26657012
rows_examined_avg: 26657012
first_seen: 2015-03-08 11:30:02
last_seen: 2015-03-08 11:30:02
digest: 7f898a15e3310816582f924af05483e0
2 rows in set (0.80 sec)
最頻繁的表
mysql> select * from schema_table_statistics limit 1\G
*************************** 1. row ***************************
table_schema: zabbix
table_name: history
total_latency: 2.10h
rows_fetched: 115738265
fetch_latency: 00:10:37.09
rows_inserted: 107732087
insert_latency: 1.68h
rows_updated: 0
update_latency: 0 ps
rows_deleted: 83830680
delete_latency: 00:14:43.18
io_read_requests: 479411
io_read: 7.31 GiB
io_read_latency: 00:01:43.52
io_write_requests: 40810690
io_write: 625.27 GiB
io_write_latency: 00:09:20.00
io_misc_requests: 876070
io_misc_latency: 22.44h
1 row in set (0.05 sec)沒有使用過的索引
mysql> select * from schema_unused_indexes limit 10;
+---------------+--------------+----------------+
| object_schema | object_name | index_name |
+---------------+--------------+----------------+
| zabbix | acknowledges | acknowledges_1 |
| zabbix | acknowledges | acknowledges_3 |
| zabbix | acknowledges | PRIMARY |
| zabbix | actions | actions_2 |
| zabbix | actions | PRIMARY |
| zabbix | alerts | alerts_1 |
| zabbix | alerts | alerts_5 |
| zabbix | alerts | alerts_2 |
| zabbix | alerts | alerts_6 |
| zabbix | alerts | PRIMARY |
+---------------+--------------+----------------+
10 rows in set (0.00 sec)可以考慮是不是真的需要這些索引
文件IO消耗
mysql> select * from io_global_by_file_by_bytes limit 3\G
*************************** 1. row ***************************
file: @@datadir/ibdata1
count_read: 691
total_read: 12.77 MiB
avg_read: 18.92 KiB
count_write: 6064498
total_written: 1.15 TiB
avg_write: 204.34 KiB
total: 1.15 TiB
write_pct: 100.00
*************************** 2. row ***************************
file: @@datadir/zabbix/history.ibd
count_read: 479383
total_read: 7.31 GiB
avg_read: 16.00 KiB
count_write: 40812693
total_written: 625.30 GiB
avg_write: 16.07 KiB
total: 632.61 GiB
write_pct: 98.84
*************************** 3. row ***************************
file: @@datadir/zabbix/history_uint.ibd
count_read: 333255
total_read: 5.09 GiB
avg_read: 16.00 KiB
count_write: 27149763
total_written: 416.14 GiB
avg_write: 16.07 KiB
total: 421.23 GiB
write_pct: 98.79
3 rows in set (0.00 sec)