Slow Query Log

The slow query log is a record of SQL statements that took a long time to perform.

A problematic SQL statement can increase the pressure on the entire cluster, resulting in a longer response time. To solve this problem, you can use the slow query log to identify the problematic statements and thus improve the performance.

Obtain the log

By grep the keyword SLOW_QUERY in the log file of TiDB, you can obtain the logs of statements whose execution time exceeds slow-threshold.

You can edit slow-threshold in the configuration file and its default value is 300ms. If you configure the slow-query-file, all the slow query logs will be written in this file.

Usage example

2018/08/20 19:52:08.632 adapter.go:363: [warning] [SLOW_QUERY] cost_time:18.647928814s
process_time:1m6.768s wait_time:12m11.212s backoff_time:600ms request_count:2058
total_keys:1869712 processed_keys:1869710 succ:true con:3 user:root@
txn_start_ts:402329674704224261 database:test table_ids:[31],index_ids:[1],
sql:select count(c) from sbtest1 use index (k_1)

Fields description

This section describes fields in the slow query log based on the usage example above.


The execution time of this statement. Only the statements whose execution time exceeds slow-threshold output this log.


The total processing time of this statement in TiKV. Because data is sent to TiKV concurrently for execution, this value might exceed cost_time.


The total waiting time of this statement in TiKV. Because the Coprocessor of TiKV runs a limited number of threads, requests might queue up when all threads of Coprocessor are working. When a request in the queue takes a long time to process, the waiting time of the subsequent requests will increase.


The waiting time before retry when this statement encounters errors that require a retry. The common errors as such include: lock occurs, Region split, the TiKV server is busy.


The number of Coprocessor requests that this statement sends.


The number of keys that Coprocessor has scanned.


The number of keys that Coprocessor has processed. Compared with total_keys, processed_keys does not include the old versions of MVCC. A great difference between processed_keys and total_keys indicates that the number of old versions are relatively large.


Whether the execution of the request succeeds or not.


Connection ID (session ID). For example, you can use the keyword con:3 to grep the log whose session ID is 3.


The name of the user who executes this statement.


The start timestamp of the transaction, that is, the ID of the transaction. You can use this value to grep the transaction-related logs.


The current database.


The IDs of the tables involved in the statement.


The IDs of the indexes involved in the statement.


The SQL statement.

Identify problematic SQL statements

Not all of the SLOW_QUERY statements are problematic. Only those whose process_time is very large will increase the pressure on the entire cluster.

The statements whose wait_time is very large and process_time is very small are usually not problematic. The large wait_time is because the statement is blocked by real problematic statements and it has to wait in the execution queue, which leads to a much longer response time.

admin show slow command

In addition to the TiDB log file, you can identify slow queries by running the admin show slow command:

admin show slow recent N
admin show slow top [internal | all] N

recent N shows the recent N slow query records, for example:

admin show slow recent 10

top N shows the slowest N query records recently (within a few days). If the internal option is provided, the returned results would be the inner SQL executed by the system; If the all option is provided, the returned results would be the user's SQL combinated with inner SQL; Otherwise, this command would only return the slow query records from the user's SQL.

admin show slow top 3
admin show slow top internal 3
admin show slow top all 5

Due to the memory footprint restriction, the stored slow query records count is limited. If the specified N is greater than the records count, the returned records count may be smaller than N.