Configure logging

edit

The logging section of the metricbeat.yml config file contains options for configuring the logging output. The logging system can write logs to the syslog or rotate log files. If logging is not explicitly configured the file output is used.

logging.level: info
logging.to_files: true
logging.files:
  path: /var/log/metricbeat
  name: metricbeat
  keepfiles: 7
  permissions: 0644

In addition to setting logging options in the config file, you can modify the logging output configuration from the command line. See Command reference.

When Metricbeat is running on a Linux system with systemd, it uses by default the -e command line option, that makes it write all the logging output to stderr so it can be captured by journald. Other outputs are disabled. See Metricbeat and systemd to know more and learn how to change this.

Configuration options

edit

You can specify the following options in the logging section of the metricbeat.yml config file:

logging.to_stderr

edit

When true, writes all logging output to standard error output. This is equivalent to using the -e command line option.

logging.to_syslog

edit

When true, writes all logging output to the syslog.

This option is not supported on Windows.

logging.to_eventlog

edit

When true, writes all logging output to the Windows Event Log.

logging.to_files

edit

When true, writes all logging output to files. The log files are automatically rotated when the log file size limit is reached.

Metricbeat only creates a log file if there is logging output. For example, if you set the log level to error and there are no errors, there will be no log file in the directory specified for logs.

logging.level

edit

Minimum log level. One of debug, info, warning, or error. The default log level is info.

debug
Logs debug messages, including a detailed printout of all events flushed. Also logs informational messages, warnings, errors, and critical errors. When the log level is debug, you can specify a list of selectors to display debug messages for specific components. If no selectors are specified, the * selector is used to display debug messages for all components.
info
Logs informational messages, including the number of events that are published. Also logs any warnings, errors, or critical errors.
warning
Logs warnings, errors, and critical errors.
error
Logs errors and critical errors.

logging.selectors

edit

The list of debugging-only selector tags used by different Metricbeat components. Use * to enable debug output for all components. Use publisher to display debug messages related to event publishing.

The list of available selectors may change between releases, so avoid creating tests that depend on specific selectors.

To see which selectors are available, run Metricbeat in debug mode (set logging.level: debug in the configuration). The selector name appears after the log level and is enclosed in brackets.

To configure multiple selectors, use the following YAML list syntax:

logging.selectors: [ harvester, input ]

To override selectors at the command line, use the -d global flag (-d also sets the debug log level). For more information, see Command reference.

logging.metrics.enabled

edit

By default, Metricbeat periodically logs its internal metrics that have changed in the last period. For each metric that changed, the delta from the value at the beginning of the period is logged. Also, the total values for all non-zero internal metrics are logged on shutdown. Set this to false to disable this behavior. The default is true.

Here is an example log line:

2017-12-17T19:17:42.667-0500    INFO    [metrics]       log/log.go:110  Non-zero metrics in the last 30s: beat.info.uptime.ms=30004 beat.memstats.gc_next=5046416

Note that we currently offer no backwards compatible guarantees for the internal metrics and for this reason they are also not documented.

logging.metrics.period

edit

The period after which to log the internal metrics. The default is 30s.

logging.metrics.namespaces

edit

A list of metrics namespaces to report in the logs. Defaults to [stats]. stats contains general Beat metrics. dataset may be present in some Beats and contains module or input metrics.

logging.files.path

edit

The directory that log files are written to. The default is the logs path. See the Directory layout section for details.

logging.files.name

edit

The name of the file that logs are written to. The default is metricbeat.

logging.files.rotateeverybytes

edit

The maximum size of a log file. If the limit is reached, a new log file is generated. The default size limit is 10485760 (10 MB).

logging.files.keepfiles

edit

The number of most recent rotated log files to keep on disk. Older files are deleted during log rotation. The default value is 7. The keepfiles options has to be in the range of 2 to 1024 files.

logging.files.permissions

edit

The permissions mask to apply when rotating log files. The default value is 0600. The permissions option must be a valid Unix-style file permissions mask expressed in octal notation. In Go, numbers in octal notation must start with 0.

Examples:

  • 0644: give read and write access to the file owner, and read access to all others.
  • 0600: give read and write access to the file owner, and no access to all others.
  • 0664: give read and write access to the file owner and members of the group associated with the file, as well as read access to all other users.

logging.files.interval

edit

Enable log file rotation on time intervals in addition to size-based rotation. Intervals must be at least 1s. Values of 1m, 1h, 24h, 7*24h, 30*24h, and 365*24h are boundary-aligned with minutes, hours, days, weeks, months, and years as reported by the local system clock. All other intervals are calculated from the unix epoch. Defaults to disabled.

logging.files.suffix

edit

When a log rotation happens it can either rename older files with an incresing index if count is configured. The other option is date that appends the current date and time to the end of the filename. When the log is rotated a new file is created and older files remain untouched.

logging.files.rotateonstartup

edit

If the log file already exists on startup, immediately rotate it and start writing to a new file instead of appending to the existing one. Defaults to true.

logging.json

edit

When true, logs messages in JSON format. The default is false.

logging.ecs

edit

When true, logs messages with minimal required Elastic Common Schema (ECS) information.

logging.files.redirect_stderr [preview] This functionality is in technical preview and may be changed or removed in a future release. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features.

edit

When true, diagnostic messages printed to Metricbeat’s standard error output will also be logged to the log file. This can be helpful in situations were Metricbeat terminates unexpectedly because an error has been detected by Go’s runtime but diagnostic information is not present in the log file. This feature is only available when logging to files (logging.to_files is true). Disabled by default.

Logging format

edit

The logging format is generally the same for each logging output. The one exception is with the syslog output where the timestamp is not included in the message because syslog adds its own timestamp.

Each log message consists of the following parts:

  • Timestamp in ISO8601 format
  • Level
  • Logger name contained in brackets (Optional)
  • File name and line number of the caller
  • Message
  • Structured data encoded in JSON (Optional)

Below are some samples:

2017-12-17T18:54:16.241-0500 INFO logp/core_test.go:13 unnamed global logger

2017-12-17T18:54:16.242-0500 INFO [example] logp/core_test.go:16 some message

2017-12-17T18:54:16.242-0500 INFO [example] logp/core_test.go:19 some message {"x": 1}