Command-Line Options for Logging

There are two different kinds of logs. Human-readable logs and machine-readable logs. The human-readable logs are used to provide an administration with information about the server. The machine-readable logs are used to provide statistics about executed requests and timings about computation steps.

General Logging Options

Logfile

--log.file filename

This option allows the user to specify the name of a file to which information is logged. By default, if no log file is specified, the standard output is used. Note that if the file named by filename does not exist, it will be created. If the file cannot be created (e.g. due to missing file privileges), the server will refuse to start. If the specified file already exists, output is appended to that file.

Use + to log to standard error. Use - to log to standard output. Use "" to disable logging to file.

--log.tty filename

Be default, if started on a tty, the log output will also go to the ttyp. Use "" to disable.

Request

--log.requests-file filename
This option allows the user to specify the name of a file to which requests are logged. By default, no log file is used and requests are not logged. Note that if the file named by filename does not exist, it will be created. If the file cannot be created (e.g. due to missing file privileges), the server will refuse to start. If the specified file already exists, output is appended to that file.
Use + to log to standard error. Use - to log to standard output. Use ”” to disable request logging altogether.
The log format is

• "http-request": static string indicating that an HTTP request was logged
• client address: IP address of client
• HTTP method type, e.g. GET, POST
• HTTP version, e.g. HTTP/1.1
• HTTP response code, e.g. 200
• request body length in bytes
• response body length in bytes
• server request processing time, containing the time span between fetching the first byte of the HTTP request and the start of the HTTP response

Human Readable Logging

Logfiles

Level

--log.level level
--log level
Allows the user to choose the level of information which is logged by the server. The argument level is specified as a string and can be one of the values listed below. Note that, fatal errors, that is, errors which cause the server to terminate, are always logged irrespective of the log level assigned by the user. The variant c log.level can be used in configuration files, the variant c log for command line options.
fatal: Logs errors which cause the server to terminate.
Fatal errors generally indicate some inconsistency with the manner in which the server has been coded. Fatal errors may also indicate a problem with the platform on which the server is running. Fatal errors always cause the server to terminate. For example,

2010-09-20T07:32:12Z [4742] FATAL a http server has already been created

error: Logs errors which the server has encountered.
These errors may not necessarily result in the termination of the server. For example,

2010-09-17T13:10:22Z [13967] ERROR strange log level 'errors'\, going to 'warning'

warning: Provides information on errors encountered by the server, which are not necessarily detrimental to it’s continued operation.
For example,

2010-09-20T08:15:26Z [5533] WARNING got corrupted HTTP request 'POS?'

Note: The setting the log level to warning will also result in all errors to be logged as well.
info: Logs information about the status of the server.
For example,

2010-09-20T07:40:38Z [4998] INFO SimpleVOC ready for business

Note: The setting the log level to info will also result in all errors and warnings to be logged as well.
debug: Logs all errors, all warnings and debug information.
Debug log information is generally useful to find out the state of the server in the case of an error. For example,

2010-09-17T13:02:53Z [13783] DEBUG opened port 7000 for any

Note: The setting the log level to debug will also result in all errors, warnings and server status information to be logged as well.
trace: As the name suggests, logs information which may be useful to trace problems encountered with using the server.
For example,

2010-09-20T08:23:12Z [5687] TRACE trying to open port 8000

Note: The setting the log level to trace will also result in all errors, warnings, status information, and debug information to be logged as well.

Local Time

--log.use-local-time
If specified, all dates and times in log messages will use the server’s local time-zone. If not specified, all dates and times in log messages will be printed in UTC / Zulu time. The date and time format used in logs is always YYYY-MM-DD HH:MM:SS, regardless of this setting. If UTC time is used, a Z will be appended to indicate Zulu time.

Line number

--log.line-number
Normally, if an human readable fatal, error, warning or info message is logged, no information about the file and line number is provided. The file and line number is only logged for debug and trace message. This option can be use to always log these pieces of information.

Prefix

--log.prefix prefix
This option is used specify an prefix to logged text.

Thread

--log.thread
Whenever log output is generated, the process ID is written as part of the log information. Setting this option appends the thread id of the calling thread to the process id. For example,

2010-09-20T13:04:01Z [19355] INFO ready for business

when no thread is logged and

2010-09-20T13:04:17Z [19371-18446744072487317056] ready for business

when this command line option is set.

Source Filter

--log.source-filter arg
For debug and trace messages, only log those messages originated from the C source file arg. The argument can be used multiple times.

Content Filter

--log.content-filter arg
Only log message containing the specified string arg.

Performance

--log.performance
If this option is set, performance-related info messages will be logged via the regular logging mechanisms. These will consist of mostly timing and debugging information for performance-critical operations.
Currently performance-related operations are logged as INFO messages. Messages starting with prefix [action] indicate that an instrumented operation was started (note that its end won’t be logged). Messages with prefix [timer] will contain timing information for operations. Note that no timing information will be logged for operations taking less time than 1 second. This is to ensure that sub-second operations do not pollute logs.
The contents of performance-related log messages enabled by this option are subject to change in future versions of ArangoDB.

Machine Readable Logging

Application

--log.application name
Specifies the name of the application which should be logged if this item of information is to be logged.

Facility

--log.facility name
If this option is set, then in addition to output being directed to the standard output (or to a specified file, in the case that the command line log.file option was set), log output is also sent to the system logging facility. The arg is the system log facility to use. See syslog for further details.
The value of arg depends on your syslog configuration. In general it will be user. Fatal messages are mapped to crit, so if arg is user, these messages will be logged as user.crit. Error messages are mapped to err. Warnings are mapped to warn. Info messages are mapped to notice. Debug messages are mapped to info. Trace messages are mapped to debug.