The slow query log records the query statements that take more time than the specified value to execute in TencentDB for SQL Server. It enables you to find out inefficient query statements for optimization. Typically, it is a SQL statement for troubleshooting and an important feature for checking the performance of the current TencentDB for SQL Server instance.
This document describes how to query and download a slow query log in the console.
Note:
The log is stored in the Beijing time zone (UTC+8) by default. If the default time zone is modified, it will be stored in the time zone of the instance server, but its storage time will still be in Beijing time in the console.
Use Limits
The collection of slow query is enabled by default and cannot be disabled.
The collection threshold of slow query is 1s (or 1,000 ms) by default, and SQL executions exceeding 1s will be recorded as a slow query log.
The slow query logs are collected every 5 minutes by default, and SQL statements exceeding 1s will be recorded.
The slow query logs are retained for 7 days by default and will be automatically deleted upon expiration.
Description of Fields Contained in Slow Query Log Files
|
client_app_name | Client application name. |
client_hostname | Client hostname. |
cpu_time | CPU time used by a remote procedure call, in microseconds. |
duration | Time required to complete a remote procedure call, in microseconds. |
last_row_count | Number of rows returned by the last executed SQL statement. |
line_number | Line number of a SQL statement in the source file. |
logical_reads | Number of logical reads during execution. |
offset | Start offset of a SQL statement in the source file. |
offset_end | End offset of a SQL statement in the source file. |
parameterized | Specifies whether a SQL statement is parameterized. |
physical_reads | Number of physical reads during execution. |
row_count | Number of rows returned. |
spills | Number of pages written to the overflow database during query execution. |
sql_text | Text content of a SQL statement. |
statement | Executed SQL statement. |
username | Username for performing an operation. |
writes | Number of writes during query execution. |
connection_reset_option | Option used to reset a connection. This field only applies to resetting the connection for remote procedure calls. |
client_connection_id | Optional identifier provided by the client upon connection. |
data_stream | Binary stream for replaying the trace of a remote procedure call. |
database_name | Database name. |
object_name | Object name referenced by a remote procedure call. |
output_parameters | Output parameters returned from a remote procedure call. |
query_plan_hash | Query plan hash, which can be used to identify similar query execution plans. You can use the query plan hash to find the cumulative overheads of queries with similar execution plans. |
result | Return value of a remote procedure call. |
Directions
2. Select the region at the top, find the target instance, and click the instance ID or Manage in the Operation column to enter the instance management page.
3. On the instance management page, select the Operation Log tab > Slow Query Log to see slow log list.
You can view the following fields: File Name, Start Time of File Generation, End Time of File Generation, File Size, and Operation (Download).
You can search for slow logs generated in the last 5/15/30 minutes, last 1/3/24 hours, today, yesterday, last 3/7/30 days, or a custom time range.
4. Click Download in the Operation column to download slow query log files.
Was this page helpful?