Question: How to find queries taking more than a given amount of time to complete?
The simplest way to find long-running queries is to look at the process list. There are various options for how to do this depending on your version and whether you have the Sys Schema installed. The examples in this post will all look for queries that have taken longer than 10 seconds. Change the duration as required.
The Sys Schema session View
In MySQL 5.6 or later with the Sys Schema installed, you can use the session view, for example:
mysql> SELECT * FROM sys.session WHERE command = 'Query' AND time > 10\G
*************************** 1. row ***************************
thd_id: 36
conn_id: 11
user: root@localhost
db: db1
command: Query
state: copy to tmp table
time: 244
current_statement: ALTER TABLE MegaCity ENGINE=InnoDB
statement_latency: 4.06 m
progress: 52.73
lock_latency: 21.55 ms
rows_examined: 0
rows_sent: 0
rows_affected: 0
tmp_tables: 0
tmp_disk_tables: 0
full_scan: NO
last_statement: NULL
last_statement_latency: NULL
current_memory: 296.29 KiB
last_wait: NULL
last_wait_latency: NULL
source: NULL
trx_latency: NULL
trx_state: NULL
trx_autocommit: NULL
pid: 11884
program_name: mysql
1 row in set (0.08 sec)
The session view is the same as the sys.processlist view with the difference that the session view excludes background threads. An advantage of the session view over the methods described below is that it will include memory usage and progress information if available. For example in the above output, the query is estimated to have completed 52.73% of the work and is currently using 296.29 KiB of memory.
Both the progress estimation and the memory usage are not exact. Particularly the memory usage depends on the enabled Performance Schema instruments and the coverage of the instruments in the code. The progress information can be used to help determine whether the query should be killed or allowed to complete.
The Performance Schema threads table
In MySQL 5.6 and later it is recommended to use the performance_schema.threads table over SHOW PROCESSLIST or the information_schema.PROCESSLIST as using the threads table has less impact on the running queries. The Sys Schema views processlist and session discussed above are built on top of the threads table. To get long running queries, you can for example use:
mysql> SELECT PROCESSLIST_ID, PROCESSLIST_USER, PROCESSLIST_HOST, PROCESSLIST_DB, PROCESSLIST_COMMAND, PROCESSLIST_TIME, PROCESSLIST_STATE, PROCESSLIST_INFO FROM performance_schema.threads WHERE PROCESSLIST_COMMAND = 'Query' AND PROCESSLIST_TIME > 10; +----------------+------------------+------------------+----------------+---------------------+------------------+---------------------------+-------------------------------------------------+ | PROCESSLIST_ID | PROCESSLIST_USER | PROCESSLIST_HOST | PROCESSLIST_DB | PROCESSLIST_COMMAND | PROCESSLIST_TIME | PROCESSLIST_STATE | PROCESSLIST_INFO | +----------------+------------------+------------------+----------------+---------------------+------------------+---------------------------+-------------------------------------------------+ | 11 | root | localhost | db1 | Query | 49 | Searching rows for update | UPDATE MegaCity SET Population = Population + 1 | +----------------+------------------+------------------+----------------+---------------------+------------------+---------------------------+-------------------------------------------------+ 1 row in set (0.00 sec)
The Information Schema PROCESSLIST table
In MySQL 5.1 and 5.5, the above solutions cannot be used. Instead, use the PROCESSLIST table in the Information Schema. Querying the PROCESSLIST table is equivalent to executing SHOW PROCESSLIST with the difference that you can specify a WHERE clause.
The Information Schema PROCESSLIST table (and SHOW PROCESSLIST) requires a mutex that can affect server performance. For this reason, it is recommended to use one of the above Performance Schema based methods in MySQL 5.6 and later and to avoid querying the PROCESSLIST table or SHOW PROCESSLIST frequently.
For example:
mysql> SELECT * FROM information_schema.PROCESSLIST WHERE Command = 'Query' AND TIME > 10; +----+------+-----------+------+---------+------+---------------------------+-------------------------------------------------+ | ID | USER | HOST | DB | COMMAND | TIME | STATE | INFO | +----+------+-----------+------+---------+------+---------------------------+-------------------------------------------------+ | 11 | root | localhost | db1 | Query | 24 | Searching rows for update | UPDATE MegaCity SET Population = Population + 1 | +----+------+-----------+------+---------+------+---------------------------+-------------------------------------------------+ 1 row in set (0.00 sec)