Log Configuration File
The validator and the Python SDK make it easy to customize Sawtooth logging output. This is done by creating a log config file in TOML or YAML format and passing it to the built-in Python logging module.
You must use YAML to configure a remote syslog service. Due to a limitation in the TOML spec, you cannot configure a remote syslog service using TOML.
About Sawtooth Log Files
If there is no log configuration file provided, Sawtooth creates an
error log and a debug log by default. These files are stored in the log
log_dir), which is
/var/log/sawtooth by default. However,
SAWTOOTH_HOME environment variable, if set, can change
the default location. For more information, see
Path Configuration File.
For Sawtooth core components, such as the validator or REST API, the log
file names are
example, the validator log files are:
For Python transaction processors, the author determines the name of the log file. Because a Sawtooth node can run several instances of the same transaction processor, it is important to create unique log file names for each running transaction processor to avoid naming conflicts.
The example transaction processors included in Sawtooth use the following naming convention:
For example, an instance of the IntegerKey (
processor could have the following log files:
To change the default logging behavior of a Sawtooth component, such as
the validator or REST API, put a log configuration file in the config
/var/log/sawtooth by default; see
Path Configuration File).
The validator log config file should be named
Sawtooth provides an example log configuration file in
/etc/sawtooth/log_config.toml.example. To create a log configuration
file, copy the example file to the config directory and name it
log_config.toml. Copy with
cp -a to preserve the file's ownership
and permissions (or change after copying to owner
sawtooth, and permissions
Each transaction processor can define its own config file. The name of this file is determined by the author. The transaction processors included in the Python SDK use the following naming convention:
For example, the IntegerKey (
intkey) log configuration file is
Example: Configure a Specific Logger
If the default logs give too much information, you can configure a specific logger that will only report on the area of the code you are interested in.
log_config.toml file creates a handler that only writes
interconnect logs to the directory and file specified.
version = 1 disable_existing_loggers = false [formatters.simple] format = "[%(asctime)s.%(msecs)03d [%(threadName)s] %(module)s %(levelname)s] %(message)s" datefmt = "%H:%M:%S" [handlers.interconnect] level = "DEBUG" formatter = "simple" class = "logging.FileHandler" filename = "path/filename.log" [loggers."sawtooth_validator.networking.interconnect"] level = "DEBUG" propagate = true handlers = [ "interconnect"]
The formatter and log level can also be specified to provide the exact information you want in your logs.
Example: Create a Rotating File Handler
This example shows how to set up rotating logs. This is useful when the logs may grow very large, such as with a long-running network. For example:
[formatters.simple] format = "[%(asctime)s.%(msecs)03d [%(threadName)s] %(module)s %(levelname)s] %(message)s" datefmt = "%H:%M:%S" [handlers.interconnect] level = "DEBUG" formatter = "simple" class = "logging.handlers.RotatingFileHandler" filename = "example-interconnect.log" maxBytes = 50000000 backupCount=20 [loggers."sawtooth_validator.networking.interconnect"] level = "DEBUG" propagate = true handlers = [ "interconnect"]
If one file exceeds the
maxBytes set in the config file, that file
will be renamed to
filename.log.1 and logs will be written to a new
filename.log. This process continues for the number of files plus one
(the value set in
backupCount). After that point, the file being
written to is rotated. The current log file is always
For more Python configuration options, see the Python documentation at https://docs.python.org/3/library/logging.config.html.