UDF Log API Overview

Logging APIs are exposed for Python UDFs. You need to import the flow_func module to use these APIs. Use the defined logger object within this module to call its encapsulated logging APIs of different severity levels.

**Table 1** Log types
Log Type	Application Scenarios	Log Level	Log Macro
Run log	Run logs record abnormal status and operations, key events, and resource usage during system running.	ERROR/INFO	Error-Level Log Macro of Run Logs/Info-Level Log Macro of Run Logs
Debug log	Debug logs record the following scenarios (including but not limited to these scenarios): - Entrance and exit of all calls such as API and function calls - Operation entrance and prerequisite setting - Start or timeout of a timer - Before and after status setting and status transfer condition judgment - Business-related resource statistics, business processing entrance/exit, and performance computing statistics - All processing failures and exceptions Debug logs record code-related information for R&D engineers to locate faults.	ERROR/WARN/INFO/DEBUG	Error-Level Log Macro of Debug Logs/Warn-Level Log Macro of Debug Logs/Error-Level Log Macro of Debug Logs/Debug-Level Log Macro of Debug Logs

Log usage example:

import dataflow.flow_func as ff
ff.logger.info("This is a test info log :%s %d %f.", "test_str", 100, 0.1)

For details about how to change the log level, see the usage of ASCEND_GLOBAL_LOG_LEVEL and ASCEND_MODULE_LOG_LEVEL in Environment Variables. User UDF logs correspond to the APP module. You can control the APP log level based on the module.

For example, if you want to enable the Info log function of a customized UDF, run the following command to enable the Info log function of the APP module:

export ASCEND_MODULE_LOG_LEVEL=APP=1

You can also run the following command to enable the Info log function for all modules:

export ASCEND_GLOBAL_LOG_LEVEL=1

During UDF execution, flow control is performed on APP module logs to prevent too many logs from affecting framework logs and execution performance. The restriction rules are as follows.

**Table 2** Restriction rules
API Type	Debug log	Debug Log	Debug Log	Debug Log	Run Log	Run Log
Level	debug	info	warn	error	run_info	run_error
DEBUG	Unlimited	Unlimited	Unlimited	Unlimited	50/400	50/400
INFO	Not print	100/1000	100/1000	100/1000	100/1000	100/1000
WARN	Not print	Not print	50/400	50/400	50/400	50/400
ERROR	Not print	Not print	Not print	50/400	50/400	50/400

Note: Each independent process outputs A logs per second, with a cumulative ceiling of B. Values are displayed as [A/B] in the table above. The cumulative ceiling refers to the maximum instantaneous log output volume available when the process has stopped printing logs for a period of time.

Parent topic: UDF log APIs