Rate This Document
Findability
Accuracy
Completeness
Readability

Tuning Result Processing Functions

The tuning result processing tool facilitates report display, single-round value setting, retests, and importance analysis in interactive mode.

Command Function

Uses the report result after automatic tuning. You can also specify the parameter file directory and perform tuning on the use screen.

Syntax

1
devkit kat use [-h] -i <dir> [-l {0,1,2,3}]

Parameter Description

Table 1 Parameter description

Parameter

Option

Description

-h/--help

-

Obtains help information.

-i/--input

-

This parameter is mandatory. It is the result directory or template file directory generated after automatic tuning.

NOTE:

If you specify the template file directory, the default values of the train subcommand parameters are the values given in the parameter configuration file because there is no historical data.

-l/--log-level

0/1/2/3

Log level, which defaults to 2.
NOTE:

The default level is 2 (WARNING).

  • 0: DEBUG
  • 1: INFO
  • 2: WARNING
  • 3: ERROR

Example

The parameter set generated after automatic tuning is located in the /opt/template/template_Postgresql_20250225_091524/train-20250225-112832 directory.

devkit kat use -i /opt/template/template_Postgresql_20250225_091524/train-20250225-112832

The following information is displayed on the interaction screen:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
...
...
...
[2025-05-20 03:12:01 UTC] [KAT] [message] - System: Initiating system parameter audit and backup process...
[2025-05-20 03:12:01 UTC] [KAT] [message] - System: System parameter backup completed successfully.

Available Commands:

    exit        - Exit the program
    help        - Show this help message
    ibesttest   - Test best parameters inferenced by parameter importance analysis
    log         - Change the log level
    man         - Manual of use
    set         - Set the config of round n
    show        - Show round <n> information
    test        - Retest round sequence. test <n> times
    train       - Additional training after original result

For more usage, please refer to "help <command>". e.g., help test
Table 2 Subcommands on the interaction screen

Subcommand

Option

Description

exit

-

Exits from the interactive screen.

help

-

Obtains help information.

ibesttest

-

Uses the optimal parameters inferred from parameter importance analysis to perform tests.

log

0/1/2/3

Log level, which defaults to 2.
  • 0: DEBUG
  • 1: INFO
  • 2: WARNING
  • 3: ERROR

man

-

Kunpeng AutoTuner manual, including the full introduction and usage examples of subcommands (show, train, test, and set).

set

-

Applies the parameter values of a tuning round to the application or system.

show

-

Displays the performance test data of some rounds and analyzes the importance of parameters.

test

-

Performs the tests again using the same parameters of some rounds.

train

-

Resumes automatic tuning. The Kunpeng AutoTuner automatically tunes task parameters based on service scenario metrics.

On the use interaction screen, the combination of round and a number indicates a specific round. For example, run show round 2 to display the information about the second tuning round.