# Logging configuration and usage

Looking for the latest Prefect 2 release? Prefect 2 and Prefect Cloud 2 have been released for General Availability. See https://docs.prefect.io/ for details.

Prefect natively ships with a default logger configured to handle the management of logs. More information on logging in Prefect can be found in the logging concept document.

# Configure the logger

The default Prefect logger can be configured by adjusting the settings found in Prefect's config.

level = "INFO"

# The log format
format = "[%(asctime)s] %(levelname)s - %(name)s | %(message)s"

# additional log attributes to extract from context
# e.g., log_attributes = "['context_var']"
log_attributes = "[]"

# the timestamp format
datefmt = "%Y-%m-%d %H:%M:%S"

# Extra loggers for Prefect log configuration
extra_loggers = "[]"


# Send flow run logs to Prefect Cloud/Server
send_flow_run_logs = false

# Logging from within tasks

The Prefect logger can be used within tasks but the way it is used depends on the API—functional or imperative—that you choose to write your tasks with.

If using the functional API to declare your tasks using the @task decorator then the logger can be instantiated by grabbing it from context:

def my_task():
    logger = prefect.context.get("logger")

    logger.info("An info message.")
    logger.warning("A warning message.")

If using the imperative API to declare your tasks as classes then the logger can be used directly from self.logger:

class MyTask(prefect.Task):
    def run(self):
        self.logger.info("An info message.")
        self.logger.warning("A warning message.")

# Logging stdout

Tasks also provide an optional argument to toggle the logging of stdout—log_stdout. This means that, if enabled, anytime you send output to stdout (such as print()) it will be logged using the Prefect logger. This option is disabled by default because this option might not be suitable for certain types of tasks.

# Logging with a backend

If you are deploying your flows with the use of a backend such as Prefect Core's server or Prefect Cloud then the above logging options apply but may not propagate to your runs. For example a flow run could have an environment variable PREFECT__LOGGING__LEVEL=DEBUG that changes the logging config for just that run.

An agent will respect the config on the machine it is started on when determining if logs should be sent to cloud. This allows you to disable sending logs to cloud per machine.


# Send flow run logs to Prefect Cloud/Server
send_flow_run_logs = false

Logging can also be disabled for flow runs by providing the --no-cloud-logs flag when starting a Prefect agent from the CLI. This will ignore the value set in the config.