path: root/lib/kernel/doc/src/logger_chapter.xml

<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">

<chapter>
  <header>
    <copyright>
      <year>2017</year><year>2018</year>
      <holder>Ericsson AB. All Rights Reserved.</holder>
    </copyright>
    <legalnotice>
      Licensed under the Apache License, Version 2.0 (the "License");
      you may not use this file except in compliance with the License.
      You may obtain a copy of the License at
 
          http://www.apache.org/licenses/LICENSE-2.0

      Unless required by applicable law or agreed to in writing, software
      distributed under the License is distributed on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
      See the License for the specific language governing permissions and
      limitations under the License.
    
    </legalnotice>

    <title>Logging</title>
    <prepared></prepared>
    <docno></docno>
    <date></date>
    <rev></rev>
    <file>logger_chapter.xml</file>
  </header>

  <p>Erlang/OTP 21.0 provides a standard API for logging
    through <c>Logger</c>, which is part of the Kernel
    application. Logger consists of the API for issuing log events,
    and a customizable backend where log handlers, filters and
    formatters can be plugged in.</p>
  <p>By default, the Kernel application installs one log handler at
    system start. This handler is named <c>default</c>. It receives
    and processes standard log events produced by the Erlang runtime
    system, standard behaviours and different Erlang/OTP
    applications. The log events are by default written to the
    terminal.</p>
  <p>You can also configure the system so that the default handler
    prints log events to a single file, or to a set of wrap logs
    via <seealso marker="disk_log"><c>disk_log</c></seealso>.</p>
  <p>By configuration, you can also modify or disable the default
    handler, replace it by a custom handler, and install additional
    handlers.</p>

  <note>
    <p>Since Logger is new in Erlang/OTP 21.0, we do reserve the right
      to introduce changes to the Logger API and functionality in
      patches following this release. These changes might or might not
      be backwards compatible with the initial version.</p>
  </note>

  <section>
    <title>Overview</title>
    <p>A <em>log event</em> consists of a <em>log level</em>, the
      <em>message</em> to be logged, and <em>metadata</em>.</p>
    <p>The Logger backend forwards log events from the API, first
      through a set of <em>primary filters</em>, then through a set of
      secondary filters attached to each log handler. The secondary
      filters are in the following named <em>handler filters</em>.</p>
    <p>Each filter set consists of a <em>log level check</em>,
      followed by zero or more <em>filter functions</em>.</p>
    <p>The following figure shows a conceptual overview of Logger. The
      figure shows two log handlers, but any number of handlers can be
      installed.</p>

    <!-- The image is edited with dia in logger_arch.dia file,
	 and .png file generated with make target 'png'. -->
    <image file="logger_arch.png">
      <icaption>Conceptual Overview</icaption>
    </image>

    <p>Log levels are expressed as atoms. Internally in Logger, the
      atoms are mapped to integer values, and a log event passes the
      log level check if the integer value of its log level is less
      than or equal to the currently configured log level. That is,
      the check passes if the event is equally or more severe than the
      configured level. See section <seealso marker="#log_level">Log
      Level</seealso> for a listing and description of all log
      levels.</p>
    <p>The primary log level can be overridden by a log level
      configured per module. This is to, for instance, allow more
      verbose logging from a specific part of the system.</p>
    <p>Filter functions can be used for more sophisticated filtering
      than the log level check provides. A filter function can stop or
      pass a log event, based on any of the event's contents. It can
      also modify all parts of the log event. See section
      <seealso marker="#filters">Filters</seealso> for more
      details.</p>
    <p>If a log event passes through all primary filters and all
      handler filters for a specific handler, Logger forwards the
      event to the <em>handler callback</em>. The handler formats and
      prints the event to its destination. See
      section <seealso marker="#handlers">Handlers</seealso> for more
      details.</p>
    <p>Everything up to and including the call to the handler
      callbacks is executed on the client process, that is, the
      process where the log event was issued. It is up to the handler
      implementation if other processes are involved or not.</p>
    <p>The handlers are called in sequence, and the order is not
      defined.</p>
  </section>
  <section>
    <marker id="logger_api"/>
    <title>Logger API</title>
    <p>The API for logging consists of a set
      of <seealso marker="logger#macros">macros</seealso>, and a set
      of functions on the form <c>logger:Level/1,2,3</c>, which are
      all shortcuts
      for <seealso marker="logger#log-2">
    <c>logger:log(Level,Arg1[,Arg2[,Arg3]])</c></seealso>.</p>
    <p>The macros are defined in <c>logger.hrl</c>, which is included
      in a module with the directive</p>
    <code>-include_lib("kernel/include/logger.hrl").</code>
    <p>The difference between using the macros and the exported
      functions is that macros add location (originator) information
      to the metadata, and performs lazy evaluation by wrapping the
      logger call in a case statement, so it is only evaluated if the
      log level of the event passes the primary log level check.</p>
    <section>
      <marker id="log_level"/>
      <title>Log Level</title>
      <p>The log level indicates the severity of a event. In
	accordance with the Syslog protocol,
	<url href="https://www.ietf.org/rfc/rfc5424.txt">RFC
	5424</url>, eight log levels can be specified. The following
	table lists all possible log levels by name (atom), integer
	value, and description:</p>

      <table align="left">
	<row>
	  <cell><strong>Level</strong></cell>
	  <cell align="center"><strong>Integer</strong></cell>
	  <cell><strong>Description</strong></cell>
	</row>
	<row>
	  <cell>emergency</cell>
	  <cell align="center">0</cell>
	  <cell>system is unusable</cell>
	</row>
	<row>
	  <cell>alert</cell>
	  <cell align="center">1</cell>
	  <cell>action must be taken immediately</cell>
	</row>
	<row>
	  <cell>critical</cell>
	  <cell align="center">2</cell>
	  <cell>critical conditions</cell>
	</row>
	<row>
	  <cell>error</cell>
	  <cell align="center">3</cell>
	  <cell>error conditions</cell>
	</row>
	<row>
	  <cell>warning</cell>
	  <cell align="center">4</cell>
	  <cell>warning conditions</cell>
	</row>
	<row>
	  <cell>notice</cell>
	  <cell align="center">5</cell>
	  <cell>normal but significant conditions</cell>
	</row>
	<row>
	  <cell>info</cell>
	  <cell align="center">6</cell>
	  <cell>informational messages</cell>
	</row>
	<row>
	  <cell>debug</cell>
	  <cell align="center">7</cell>
	  <cell>debug-level messages</cell>
	</row>
	<tcaption>Log Levels</tcaption>
      </table>
      <p>Notice that the integer value is only used internally in
	Logger. In the API, you must always use the atom. To compare
	the severity of two log levels,
	use <seealso marker="logger#compare_levels-2">
	  <c>logger:compare_levels/2</c></seealso>.</p>
    </section>
    <section>
      <marker id="log_message"/>
      <title>Log Message</title>
      <p>The log message contains the information to be logged. The
	message can consist of a format string and arguments (given as
	two separate parameters in the Logger API), a string or a
	report. The latter, which is either a map or a key-value list,
	can be accompanied by a <em>report callback</em> specified in
	the log event's <seealso marker="#metadata">metadata</seealso>.
	The report callback is a convenience function that
	the <seealso marker="#formatters">formatter</seealso> can use
	to convert the report to a format string and arguments, or
	directly to a string. The
	formatter can also use its own conversion function, if no
	callback is provided, or if a customized formatting is
	desired.</p>
      <p>The report callback must be a fun with one or two
	arguments. If it takes one argument, this is the report
	itself, and the fun returns a format string and arguments:</p>
      <pre>fun((<seealso marker="logger#type-report"><c>logger:report()</c></seealso>) -> {<seealso marker="stdlib:io#type-format"><c>io:format()</c></seealso>,[term()]})</pre>
      <p>If it takes two arguments, the first is the report, and the
	second is a map containing extra data that allows direct
	coversion to a string:</p>
      <pre>fun((<seealso marker="logger#type-report"><c>logger:report()</c></seealso>,<seealso marker="logger#type-report_cb_config"><c>logger:report_cb_config()</c></seealso>) -> <seealso marker="stdlib:unicode#type-chardata"><c>unicode:chardata()</c></seealso>)
      </pre>
      <p>The fun must obey the <c>depth</c> and <c>chars_limit</c>
	parameters provided in the second argument, as the formatter
	cannot do anything useful of these parameters with the
	returned string. The extra data also contains a field named
	<c>single_line</c>, indicating if the printed log message may
	contain line breaks or not. This variant is used when the
	formatting of the report depends on the size or single line
	parameters.</p>
      <p>Example, format string and arguments:</p>
      <code>logger:error("The file does not exist: ~ts",[Filename])</code>
      <p>Example, string:</p>
      <code>logger:notice("Something strange happened!")</code>
      <p>Example, report, and metadata with report callback:</p>
      <code>
logger:debug(#{got => connection_request, id => Id, state => State},
             #{report_cb => fun(R) -> {"~p",[R]} end})</code>
      <p>The log message can also be provided through a fun for lazy
	evaluation. The fun is only evaluated if the primary log level
	check passes, and is therefore recommended if it is expensive
	to generate the message. The lazy fun must return a string, a
	report, or a tuple with format string and arguments.</p>
    </section>
    <section>
      <title>Metadata</title>
      <p>Metadata contains additional data associated with a log
	message. Logger inserts some metadata fields by default, and
	the client can add custom metadata in two different ways:</p>
      <taglist>
	<tag>Set process metadata</tag>
	<item>
	  <p>Process metadata is set and updated
	    with <seealso marker="logger#set_process_metadata-1">
	      <c>logger:set_process_metadata/1</c></seealso>
	    and <seealso marker="logger#update_process_metadata-1">
	      <c>logger:update_process_metadata/1</c></seealso>,
	    respectively. This metadata applies to the process on
	    which these calls are made, and Logger adds the metadata
	    to all log events issued on that process.</p>
	</item>
	<tag>Add metadata to a specific log event</tag>
	<item>
	  <p>Metadata associated with one specific log event is given
	    as the last parameter to the log macro or Logger API
	    function when the event is issued. For example:</p>
	  <code>?LOG_ERROR("Connection closed",#{context => server})</code>
	</item>
      </taglist>
      <p>See the description of
	the <seealso marker="logger#type-metadata">
	  <c>logger:metadata()</c></seealso> type for information
	about which default keys Logger inserts, and how the different
	metadata maps are merged.</p>
    </section>
  </section>
  <section>
    <marker id="filter"/>
    <title>Filters</title>
    <p>Filters can be primary, or attached to a specific
      handler. Logger calls the primary filters first, and if they all
      pass, it calls the handler filters for each handler. Logger
      calls the handler callback only if all filters attached to the
      handler in question also pass.</p>
    <p>A filter is defined as:</p>
    <pre>{FilterFun, Extra}</pre>
    <p>where <c>FilterFun</c> is a function of arity 2,
      and <c>Extra</c> is any term. When applying the filter, Logger
      calls the function with the log event as the first argument,
      and the value of <c>Extra</c> as the second
      argument. See <seealso marker="logger#type-filter">
	<c>logger:filter()</c></seealso> for type definitions.</p>
    <p>The filter function can return <c>stop</c>, <c>ignore</c> or
      the (possibly modified) log event.</p>
    <p>If <c>stop</c> is returned, the log event is immediately
      discarded. If the filter is primary, no handler filters or
      callbacks are called. If it is a handler filter, the
      corresponding handler callback is not called, but the log event
      is forwarded to filters attached to the next handler, if
      any.</p>
    <p>If the log event is returned, the next filter function is
      called with the returned value as the first argument. That is,
      if a filter function modifies the log event, the next filter
      function receives the modified event. The value returned from
      the last filter function is the value that the handler callback
      receives.</p>
    <p>If the filter function returns <c>ignore</c>, it means that it
      did not recognize the log event, and thus leaves to other
      filters to decide the event's destiny.</p>
    <p>The configuration option <c>filter_default</c> specifies the
      behaviour if all filter functions return <c>ignore</c>, or if no
      filters exist. <c>filter_default</c> is by default set
      to <c>log</c>, meaning that if all existing filters ignore a log
      event, Logger forwards the event to the handler
      callback. If <c>filter_default</c> is set to <c>stop</c>, Logger
      discards such events.</p>
    <p>Primary filters are added
      with <seealso marker="logger#add_primary_filter-2">
	<c>logger:add_primary_filter/2</c></seealso>
      and removed
      with <seealso marker="logger#remove_primary_filter-1">
	<c>logger:remove_primary_filter/1</c></seealso>. They can also
      be added at system start via the Kernel configuration
      parameter <seealso marker="#logger_parameter"><c>logger</c></seealso>.</p>
    <p>Handler filters are added
      with <seealso marker="logger#add_handler_filter-3">
	<c>logger:add_handler_filter/3</c></seealso>
      and removed
      with <seealso marker="logger#remove_handler_filter-2">
	<c>logger:remove_handler_filter/2</c></seealso>. They can also
      be specified directly in the configuration when adding a handler
      with <seealso marker="logger#add_handler/3">
	<c>logger:add_handler/3</c></seealso>
      or via the Kernel configuration
      parameter <seealso marker="#logger_parameter"><c>logger</c></seealso>.</p>

    <p>To see which filters are currently installed in the system,
      use <seealso marker="logger#get_config-0">
	<c>logger:get_config/0</c></seealso>,
      or <seealso marker="logger#get_primary_config-0">
	<c>logger:get_primary_config/0</c></seealso>
      and <seealso marker="logger#get_handler_config-1">
	<c>logger:get_handler_config/1</c></seealso>. Filters are
	listed in the order they are applied, that is, the first
	filter in the list is applied first, and so on.</p>

    <p>For convenience, the following built-in filters exist:</p>

    <taglist>
	<tag><seealso marker="logger_filters#domain-2">
	    <c>logger_filters:domain/2</c></seealso></tag>
	<item>
	  <p>Provides a way of filtering log events based on a
	  <c>domain</c> field in <c>Metadata</c>.</p>
	</item>
	<tag><seealso marker="logger_filters#level-2">
	  <c>logger_filters:level/2</c></seealso></tag>
	<item>
	  <p>Provides a way of filtering log events based on the log
	    level.</p>
	</item>
	<tag><seealso marker="logger_filters#progress-2">
	    <c>logger_filters:progress/2</c></seealso></tag>
	<item>
	  <p>Stops or allows progress reports from <c>supervisor</c>
	    and <c>application_controller</c>.</p>
	</item>
	<tag><seealso marker="logger_filters#remote_gl-2">
	    <c>logger_filters:remote_gl/2</c></seealso></tag>
	<item>
	  <p>Stops or allows log events originating from a process
	    that has its group leader on a remote node.</p>
	</item>
    </taglist>
  </section>

  <section>
    <marker id="handlers"/>
    <title>Handlers</title>
    <p>A handler is defined as a module exporting at least the
      following callback function:</p>

    <pre><seealso marker="logger#HModule:log-2">log(LogEvent, Config) -> void()</seealso></pre>

    <p>This function is called when a log event has passed through all
      primary filters, and all handler filters attached to the handler
      in question. The function call is executed on the client
      process, and it is up to the handler implementation if other
      processes are involved or not.</p>

    <p>Logger allows adding multiple instances of a handler
      callback. That is, if a callback module implementation allows
      it, you can add multiple handler instances using the same
      callback module. The different instances are identified by
      unique handler identities.</p>

    <p>In addition to the mandatory callback function <c>log/2</c>, a
      handler module can export the optional callback
      functions <c>adding_handler/1</c>, <c>changing_config/3</c>,
      <c>filter_config/1</c>, and <c>removing_handler/1</c>. See
      section <seealso marker="logger#handler_callback_functions">Handler
      Callback Functions</seealso> in the logger(3) manual page for
      more information about these function.</p>

    <p>The following built-in handlers exist:</p>

    <taglist>
      <tag><c>logger_std_h</c></tag>
      <item>
	<p>This is the default handler used by OTP. Multiple instances
	  can be started, and each instance will write log events to a
	  given destination, terminal or file.</p>
      </item>

      <tag><c>logger_disk_log_h</c></tag>
      <item>
	<p>This handler behaves much like <c>logger_std_h</c>, except it uses
	  <seealso marker="disk_log"><c>disk_log</c></seealso> as its
	  destination.</p>
      </item>

      <tag><marker id="ErrorLoggerManager"/><c>error_logger</c></tag>
      <item>
	<p>This handler is provided for backwards compatibility
	  only. It is not started by default, but will be
	  automatically started the first time an <c>error_logger</c>
	  event handler is added
	  with <seealso marker="error_logger#add_report_handler-1">
	    <c>error_logger:add_report_handler/1,2</c></seealso>.</p>

	<p>The old <c>error_logger</c> event handlers in STDLIB and
	  SASL still exist, but they are not added by Erlang/OTP 21.0
	  or later.</p>
      </item>
    </taglist>
  </section>

  <section>
    <marker id="formatters"/>
    <title>Formatters</title>
    <p>A formatter can be used by the handler implementation to do the
      final formatting of a log event, before printing to the
      handler's destination. The handler callback receives the
      formatter information as part of the handler configuration,
      which is passed as the second argument
      to <seealso marker="logger#HModule:log-2">
	<c>HModule:log/2</c></seealso>.</p>
    <p>The formatter information consist of a formatter
      module, <c>FModule</c> and its
      configuration, <c>FConfig</c>. <c>FModule</c> must export the
      following function, which can be called by the handler:</p>
    <pre><seealso marker="logger#FModule:format-2">format(LogEvent,FConfig)
	-> FormattedLogEntry</seealso></pre>
    <p>The formatter information for a handler is set as a part of its
      configuration when the handler is added. It can also be changed
      during runtime
      with <seealso marker="logger#set_handler_config-3">
	<c>logger:set_handler_config(HandlerId,formatter,{FModule,FConfig})</c>
      </seealso>, which overwrites the current formatter information,
      or with <seealso marker="logger#update_formatter_config-2">
	<c>logger:update_formatter_config/2,3</c></seealso>, which
	only modifies the formatter configuration.</p>
    <p>If the formatter module exports the optional callback
      function <seealso marker="logger#FModule:check_config-1">
	<c>check_config(FConfig)</c></seealso>, Logger calls this
      function when the formatter information is set or modified, to
      verify the validity of the formatter configuration.</p>
    <p>If no formatter information is specified for a handler, Logger
      uses <c>logger_formatter</c> as default. See
      the <seealso marker="logger_formatter"><c>logger_formatter(3)</c></seealso>
      manual page for more information about this module.</p>
  </section>

  <section>
    <title>Configuration</title>

    <p>At system start, Logger is configured through Kernel
      configuration parameters. The parameters that apply to Logger
      are described in
      section <seealso marker="#kernel_config_params">Kernel
      Configuration Parameters</seealso>. Examples are found in
      section <seealso marker="#config_examples">Configuration
      Examples</seealso>.</p>
    <p>During runtime, Logger configuration is changed via API
      functions. See
      section <seealso marker="logger#configuration_API">Configuration
      API Functions</seealso> in the <c>logger(3)</c> manual page.</p>

    <section>
      <title>Primary Logger Configuration</title>
      <p>Logger API functions that apply to the primary Logger
	configuration are:</p>
      <list>
	<item><seealso marker="logger#get_primary_config-0">
	    <c>get_primary_config/0</c></seealso></item>
	<item><seealso marker="logger#set_primary_config-1">
	    <c>set_primary_config/1,2</c></seealso></item>
	<item><seealso marker="logger#update_primary_config-1">
	    <c>update_primary_config/1</c></seealso></item>
	<item><seealso marker="logger#add_primary_filter-2">
	    <c>add_primary_filter/2</c></seealso></item>
	<item><seealso marker="logger#remove_primary_filter-1">
	    <c>remove_primary_filter/1</c></seealso></item>
      </list>
      <p>The primary Logger configuration is a map with the following
	keys:</p>
      <taglist>
	<tag><marker id="primary_level"/>
	  <c>level = </c><seealso marker="logger#type-level">
	    <c>logger:level()</c></seealso><c> | all | none</c></tag>
	<item>
	  <p>Specifies the primary log level, that is, log event that
	    are equally or more severe than this level, are forwarded
	    to the primary filters. Less severe log events are
	    immediately discarded.</p>
	  <p>See section <seealso marker="#log_level">Log
	      Level</seealso> for a listing and description of
	      possible log levels.</p>
	  <p>The initial value of this option is set by the Kernel
	    configuration parameter <seealso marker="#logger_level">
	      <c>logger_level</c></seealso>. It is changed during
	    runtime with <seealso marker="logger#set_primary_config-2">
	      <c>logger:set_primary_config(level,Level)</c></seealso>.</p>
	  <p>Defaults to <c>notice</c>.</p>
	</item>
	<tag><c>filters = [{FilterId,Filter}]</c></tag>
	<item>
	  <p>Specifies the primary filters.</p>
	  <list>
	    <item><c>FilterId = </c><seealso marker="logger#type-filter_id">
		<c>logger:filter_id()</c></seealso></item>
	    <item><c>Filter = </c><seealso marker="logger#type-filter">
		<c>logger:filter()</c></seealso></item>
	  </list>
	  <p>The initial value of this option is set by the Kernel
	    configuration
	    parameter <seealso marker="#logger_parameter"><c>logger</c></seealso>.
	    During runtime, primary filters are added and removed with
	    <seealso marker="logger#add_primary_filter-2">
	      <c>logger:add_primary_filter/2</c></seealso> and
	    <seealso marker="logger#remove_primary_filter-1">
	      <c>logger:remove_primary_filter/1</c></seealso>,
	    respectively.</p>
	  <p>See section <seealso marker="#filters">Filters</seealso>
	    for more detailed information.</p>
	  <p>Defaults to <c>[]</c>.</p>
	</item>
	<tag><c>filter_default = log | stop</c></tag>
	<item>
	  <p>Specifies what happens to a log event if all filters
	    return <c>ignore</c>, or if no filters exist.</p>
	  <p>See section <seealso marker="#filters">Filters</seealso>
	    for more information about how this option is used.</p>
	  <p>Defaults to <c>log</c>.</p>
	</item>
      </taglist>
    </section>

    <section>
      <marker id="handler_configuration"/>
      <title>Handler Configuration</title>
      <p>Logger API functions that apply to handler configuration
	are:</p>
      <list>
	<item><seealso marker="logger#get_handler_config-0">
	    <c>get_handler_config/0,1</c></seealso></item>
	<item><seealso marker="logger#set_handler_config-2">
	    <c>set_handler_config/2,3</c></seealso></item>
	<item><seealso marker="logger#update_handler_config-2">
	    <c>update_handler_config/2,3</c></seealso></item>
	<item><seealso marker="logger#add_handler_filter-3">
	    <c>add_handler_filter/3</c></seealso></item>
	<item><seealso marker="logger#remove_handler_filter-2">
	    <c>remove_handler_filter/2</c></seealso></item>
	<item><seealso marker="logger#update_formatter_config-2">
	    <c>update_formatter_config/2,3</c></seealso></item>
      </list>
      <p>The configuration for a handler is a map with the following keys:</p>
      <taglist>
	<tag><c>id = </c><seealso marker="logger#type-handler_id">
	    <c>logger:handler_id()</c></seealso></tag>
	<item>
	  <p>Automatically inserted by Logger. The value is the same
	    as the <c>HandlerId</c> specified when adding the handler,
	    and it cannot be changed.</p>
	</item>
	<tag><c>module = module()</c></tag>
	<item>
	  <p>Automatically inserted by Logger. The value is the same
	    as the <c>Module</c> specified when adding the handler,
	    and it cannot be changed.</p>
	</item>
	<tag><c>level = </c><seealso marker="logger#type-level">
	    <c>logger:level()</c></seealso><c> | all | none</c></tag>
	<item>
	  <p>Specifies the log level for the handler, that is, log
	    events that are equally or more severe than this level,
	    are forwarded to the handler filters for this
	    handler.</p>
	  <p>See section <seealso marker="#log_level">Log
	      Level</seealso> for a listing and description of
	      possible log levels.</p>
	  <p>The log level is specified when adding the handler, or
	    changed during runtime with, for
	    instance, <seealso marker="logger#set_handler_config/3">
	      <c>logger:set_handler_config(HandlerId,level,Level)</c></seealso>.
	  </p>
	  <p>Defaults to <c>all</c>.</p>
	</item>
	<tag><c>filters = [{FilterId,Filter}]</c></tag>
	<item>
	  <p>Specifies the handler filters.</p>
	  <list>
	    <item><c>FilterId = </c><seealso marker="logger#type-filter_id">
		<c>logger:filter_id()</c></seealso></item>
	    <item><c>Filter = </c><seealso marker="logger#type-filter">
		<c>logger:filter()</c></seealso></item>
	  </list>
	  <p>Handler filters are specified when adding the handler,
	    or added or removed during runtime with
	    <seealso marker="logger#add_handler_filter-3">
	      <c>logger:add_handler_filter/3</c></seealso> and
	    <seealso marker="logger#remove_handler_filter-2">
	      <c>logger:remove_handler_filter/2</c></seealso>,
	    respectively.</p>
	  <p>See <seealso marker="#filters">Filters</seealso> for more
	    detailed information.</p>
	  <p>Defaults to <c>[]</c>.</p>
	</item>
	<tag><c>filter_default = log | stop</c></tag>
	<item>
	  <p>Specifies what happens to a log event if all filters
	    return <c>ignore</c>, or if no filters exist.</p>
	  <p>See section <seealso marker="#filters">Filters</seealso>
	    for more information about how this option is used.</p>
	  <p>Defaults to <c>log</c>.</p>
	</item>
	<tag><c>formatter = {FormatterModule,FormatterConfig}</c></tag>
	<item>
	  <p>Specifies a formatter that the handler can use for
	    converting the log event term to a printable string.</p>
	  <list>
	    <item><c>FormatterModule = module()</c></item>
	    <item><c>FormatterConfig = </c>
		<seealso marker="logger#type-formatter_config">
		  <c>logger:formatter_config()</c></seealso></item>
	    </list>
	  <p>The formatter information is specified when adding the
	    handler. The formatter configuration can be changed during
	    runtime
	    with <seealso marker="logger#update_formatter_config-2">
	      <c>logger:update_formatter_config/2,3</c></seealso>,
	    or the complete formatter information can be overwritten
	    with, for
	    instance, <seealso marker="logger#set_handler_config-3">
	      <c>logger:set_handler_config/3</c></seealso>.</p>
	  <p>See
	    section <seealso marker="#formatters">Formatters</seealso>
	    for more detailed information.</p>
	  <p>Defaults
	    to <c>{logger_formatter,DefaultFormatterConfig}</c>. See
	    the <seealso marker="logger_formatter">
	      <c>logger_formatter(3)</c></seealso> manual page for
	    information about this formatter and its default
	    configuration.</p>
	</item>
	<tag><c>config = term()</c></tag>
	<item>
	  <p>Handler specific configuration, that is, configuration
	    data related to a specific handler implementation.</p>
	  <p>The configuration for the built-in handlers is described
            in
            the <seealso marker="logger_std_h"><c>logger_std_h(3)</c></seealso>
            and
            <seealso marker="logger_disk_log_h"><c>logger_disk_log_h(3)</c>
	    </seealso> manual pages.</p>
	</item>
      </taglist>

      <p>Notice that <c>level</c> and <c>filters</c> are obeyed by
	Logger itself before forwarding the log events to each
	handler, while <c>formatter</c> and all handler specific
	options are left to the handler implementation.</p>
    </section>

    <section>
      <marker id="kernel_config_params"/>
      <title>Kernel Configuration Parameters</title>

    <p>The following Kernel configuration parameters apply to
      Logger:</p>
    <taglist>
      <tag><marker id="logger_parameter"/><c>logger = [Config]</c></tag>
      <item>
        <p>Specifies the configuration
	  for <seealso marker="logger">Logger</seealso>, except the
	  primary log level, which is specified
	  with <seealso marker="#logger_level"><c>logger_level</c></seealso>,
	  and the compatibility
	  with <seealso marker="sasl:error_logging">SASL Error
	  Logging</seealso>, which is specified
	  with <seealso marker="#logger_sasl_compatible">
	    <c>logger_sasl_compatible</c></seealso>.</p>
	<p>With this parameter, you can modify or disable the default
	  handler, add custom handlers and primary logger filters, set
	  log levels per module, and modify
	  the <seealso marker="#proxy">proxy</seealso>
	  configuration.</p>
	<p><c>Config</c> is any (zero or more) of the following:</p>
	<taglist>
          <tag><c>{handler, default, undefined}</c></tag>
          <item>
	    <p>Disables the default handler. This allows another
              application to add its own default handler.</p>
	    <p>Only one entry of this type is allowed.</p>
	  </item>
          <tag><c>{handler, HandlerId, Module, HandlerConfig}</c></tag>
          <item>
	    <p>If <c>HandlerId</c> is <c>default</c>, then this entry
	      modifies the default handler, equivalent to calling</p>
	    <pre><seealso marker="logger#remove_handler-1">
		logger:remove_handler(default)
	    </seealso></pre>
	    <p>followed by</p>
	    <pre><seealso marker="logger#add_handler-3">
		logger:add_handler(default, Module, HandlerConfig)
	    </seealso></pre>
	    <p>For all other values of <c>HandlerId</c>, this entry
	      adds a new handler, equivalent to calling</p>
	    <pre><seealso marker="logger#add_handler/3">
		logger:add_handler(HandlerId, Module, HandlerConfig)
	    </seealso></pre>
	    <p>Multiple entries of this type are allowed.</p></item>
          <tag><c>{filters, FilterDefault, [Filter]}</c></tag>
          <item>
	    <p>Adds the specified primary filters.</p>
	    <list>
              <item><c>FilterDefault = log | stop</c></item>
              <item><c>Filter = {FilterId, {FilterFun, FilterConfig}}</c></item>
	    </list>
	    <p>Equivalent to calling</p>
	    <pre><seealso marker="logger#add_primary_filter/2">
		logger:add_primary_filter(FilterId, {FilterFun, FilterConfig})
	    </seealso></pre>
	    <p>for each <c>Filter</c>.</p>
	    <p><c>FilterDefault</c> specifies the behaviour if all
	      primary filters return <c>ignore</c>, see
	      section <seealso marker="#filters">Filters</seealso>.</p>
	    <p>Only one entry of this type is allowed.</p>
	  </item>
          <tag><c>{module_level, Level, [Module]}</c></tag>
          <item>
	    <p>Sets module log level for the given modules. Equivalent
	      to calling</p>
	    <pre><seealso marker="logger#set_module_level/2">
		logger:set_module_level(Module, Level)</seealso></pre>
	    <p>for each <c>Module</c>.</p>
	    <p>Multiple entries of this type are allowed.</p>
	  </item>
	  <tag><c>{proxy, ProxyConfig}</c></tag>
	  <item>
	    <p>Sets the proxy configuration, equivalent to calling</p>
	    <pre><seealso marker="logger#set_proxy_config/1">
		logger:set_proxy_config(ProxyConfig)
	    </seealso></pre>
	    <p>Only one entry of this type is allowed.</p>
	  </item>
	</taglist>
	<p>See
	  section <seealso marker="#config_examples">Configuration
	  Examples</seealso> for examples using the <c>logger</c>
	  parameter for system configuration.</p>
      </item>
      <tag><marker id="logger_level"/>
	<c>logger_level = Level</c></tag>
      <item>
	<p>Specifies the primary log level. See
	  the <seealso marker="kernel_app#logger_level"><c>kernel(6)</c></seealso>
	  manual page for more information about this parameter.</p>
      </item>
      <tag><marker id="logger_sasl_compatible"/>
	<c>logger_sasl_compatible = true | false</c></tag>
      <item>
	<p>Specifies Logger's compatibility
	  with <seealso marker="sasl:error_logging">SASL Error
	    Logging</seealso>. See
	  the <seealso marker="kernel_app#logger_sasl_compatible">
	    <c>kernel(6)</c></seealso> manual page for more
	  information about this parameter.</p>
      </item>
    </taglist>
    </section>

    <section>
      <marker id="config_examples"/>
      <title>Configuration Examples</title>
      <p>The value of the Kernel configuration parameter <c>logger</c>
	is a list of tuples. It is possible to write the term on the
	command line when starting an erlang node, but as the term
	grows, a better approach is to use the system configuration
	file. See
	the <seealso marker="config"><c>config(4)</c></seealso> manual
	page for more information about this file.</p>
      <p>Each of the following examples shows a simple system
	configuration file that configures Logger according to the
	description.</p>
      <p>Modify the default handler to print to a file instead of
	<c>standard_io</c>:</p>
      <code>
[{kernel,
  [{logger,
    [{handler, default, logger_std_h,  % {handler, HandlerId, Module,
      #{config => #{file => "log/erlang.log"}}}  % Config}
    ]}]}].
      </code>
      <p>Modify the default handler to print each log event as a
	single line:</p>
      <code>
[{kernel,
  [{logger,
    [{handler, default, logger_std_h,
      #{formatter => {logger_formatter, #{single_line => true}}}}
    ]}]}].
      </code>
      <p>Modify the default handler to print the pid of the logging
	process for each log event:</p>
      <code>
[{kernel,
  [{logger,
    [{handler, default, logger_std_h,
      #{formatter => {logger_formatter,
                        #{template => [time," ",pid," ",msg,"\n"]}}}}
    ]}]}].
      </code>
      <p>Modify the default handler to only print errors and more
	severe log events to "log/erlang.log", and add another handler
	to print all log events to "log/debug.log".</p>
      <code>
[{kernel,
  [{logger,
    [{handler, default, logger_std_h,
      #{level => error,
        config => #{file => "log/erlang.log"}}},
     {handler, info, logger_std_h,
      #{level => debug,
        config => #{file => "log/debug.log"}}}
    ]}]}].
      </code>
    </section>

  </section>

  <section>
    <marker id="compatibility"/>
    <title>Backwards Compatibility with error_logger</title>
    <p>Logger provides backwards compatibility with
      <c>error_logger</c> in the following ways:</p>

    <taglist>
      <tag>API for Logging</tag>
      <item>
	<p>The <c>error_logger</c> API still exists, but should only
	  be used by legacy code. It will be removed in a later
	  release.</p>
	<p>Calls
	  to <seealso marker="error_logger#error_report-1">
	    <c>error_logger:error_report/1,2</c></seealso>,
	  <seealso marker="error_logger#error_msg-1">
	    <c>error_logger:error_msg/1,2</c></seealso>, and
	  corresponding functions for warning and info messages, are
	  all forwarded to Logger as calls
	  to <seealso marker="logger#log-3">
	    <c>logger:log(Level,Report,Metadata)</c></seealso>.</p>
	<p><c>Level = error | warning | info</c> and is taken
	  from the function name. <c>Report</c> contains the actual
	  log message, and <c>Metadata</c> contains additional
	  information which can be used for creating backwards
	  compatible events for legacy <c>error_logger</c> event
	  handlers, see
	  section <seealso marker="#legacy_event_handlers">Legacy
	  Event Handlers</seealso>.</p>
      </item>
      <tag>Output Format</tag>
      <item>
	<p>To get log events on the same format as produced
	  by <c>error_logger_tty_h</c> and <c>error_logger_file_h</c>,
	  use the default formatter, <c>logger_formatter</c>, with
	  configuration parameter <c>legacy_header</c> set
	  to <c>true</c>. This is the default configuration of
	  the <c>default</c> handler started by Kernel.</p>
      </item>
      <tag>Default Format of Log Events from OTP</tag>
      <item>
	<p>By default, all log events originating from within OTP,
	  except the former so called "SASL reports", look the same as
	  before.</p>
      </item>
      <tag><marker id="sasl_reports"/>SASL Reports</tag>
      <item>
	<p>By SASL reports we mean supervisor reports, crash reports
	  and progress reports.</p>
	<p>Prior to Erlang/OTP 21.0, these reports were only logged
	  when the SASL application was running, and they were printed
	  trough SASL's own event handlers <c>sasl_report_tty_h</c>
	  and <c>sasl_report_file_h</c>.</p>
	<p>The destination of these log events was configured by
	  <seealso marker="sasl:sasl_app#deprecated_error_logger_config">SASL
	  configuration parameters</seealso>.</p>
	<p>Due to the specific event handlers, the output format
	  slightly differed from other log events.</p>
	<p>As of Erlang/OTP 21.0, the concept of SASL reports is
	  removed, meaning that the default behaviour is as
	  follows:</p>
	<list>
	  <item>Supervisor reports, crash reports, and progress reports
	    are no longer connected to the SASL application.</item>
	  <item>Supervisor reports and crash reports are issued
	    as <c>error</c> level log events, and are logged through
	    the default handler started by Kernel.</item>
	  <item>Progress reports are issued as <c>info</c> level log
	    events, and since the default primary log level
	    is <c>notice</c>, these are not logged by default. To
	    enable printing of progress reports, set
	    the <seealso marker="#primary_level">primary log
	    level</seealso> to <c>info</c>.</item>
	  <item>The output format is the same for all log
	    events.</item>
	</list>
	<p>If the old behaviour is preferred, the Kernel configuration
	  parameter <seealso marker="kernel_app#logger_sasl_compatible">
	    <c>logger_sasl_compatible</c></seealso> can be set
	  to <c>true</c>. The
	  <seealso marker="sasl:sasl_app#deprecated_error_logger_config">SASL
	  configuration parameters</seealso> can then be used as
	  before, and the SASL reports will only be printed if the
	  SASL application is running, through a second log handler
	  named <c>sasl</c>.</p>
	<p>All SASL reports have a metadata field <c>domain</c> which
	  is set to <c>[otp,sasl]</c>. This field can be
	  used by filters to stop or allow the log events.</p>
	<p>See section <seealso marker="sasl:error_logging">SASL User's
	  Guide</seealso> for more information about the old SASL
	  error logging functionality.</p>
      </item>
      <tag><marker id="legacy_event_handlers"/>Legacy Event Handlers</tag>
      <item>
	<p>To use event handlers written for <c>error_logger</c>, just
	  add your event handler with</p>
	<code>
error_logger:add_report_handler/1,2.
	</code>
	<p>This automatically starts the error logger event manager,
	  and adds <c>error_logger</c> as a handler to Logger, with
	  the following configuration:</p>
<code>
#{level => info,
  filter_default => log,
  filters => []}.
</code>
	<note>
	  <p>This handler ignores events that do not originate from
	    the <c>error_logger</c> API, or from within OTP. This
	    means that if your code uses the Logger API for logging,
	    then your log events will be discarded by this
	    handler.</p>
	  <p>The handler is not overload protected.</p>
	</note>
      </item>
    </taglist>
  </section>


  <section>
    <title>Error Handling</title>
    <p>Logger does, to a certain extent, check its input data before
      forwarding a log event to filters and handlers. It does,
      however, not evaluate report callbacks, or check the validity of
      format strings and arguments. This means that all filters and
      handlers must be careful when formatting the data of a log
      event, making sure that it does not crash due to bad input data
      or faulty callbacks.</p>
    <p>If a filter or handler still crashes, Logger will remove the
      filter or handler in question from the configuration, and print
      a short error message to the terminal. A debug event containing
      the crash reason and other details is also issued.</p>
    <p>See section <seealso marker="#log_message">Log
      Message</seealso> for more information about report callbacks
      and valid forms of log messages.</p>
  </section>

  <section>
    <title>Example: Add a handler to log info events to file</title>
    <p>When starting an Erlang node, the default behaviour is that all
      log events on level <c>notice</c> or more severe, are logged to
      the terminal via the default handler. To also log info events,
      you can either change the primary log level to <c>info</c>:</p>
    <pre>
1> <input>logger:set_primary_config(level, info).</input>
ok</pre>
    <p>or set the level for one or a few modules only:</p>
    <pre>
2> <input>logger:set_module_level(mymodule, info).</input>
ok</pre>
    <p>This allows info events to pass through to the default handler,
      and be printed to the terminal as well. If there are many info
      events, it can be useful to print these to a file instead.</p>
    <p>First, set the log level of the default handler
      to <c>notice</c>, preventing it from printing info events to the
      terminal:</p>
    <pre>
3> <input>logger:set_handler_config(default, level, notice).</input>
ok</pre>
    <p>Then, add a new handler which prints to file. You can use the
      handler
      module <seealso marker="logger_std_h"><c>logger_std_h</c></seealso>,
      and configure it to log to file:</p>
    <pre>
4> <input>Config = #{config => #{file => "./info.log"}, level => info}.</input>
#{config => #{file => "./info.log"},level => info}
5> <input>logger:add_handler(myhandler, logger_std_h, Config).</input>
ok</pre>
    <p>Since <c>filter_default</c> defaults to <c>log</c>, this
      handler now receives all log events. If you want info events
      only in the file, you must add a filter to stop all non-info
      events. The built-in
      filter <seealso marker="logger_filters#level-2">
	<c>logger_filters:level/2</c></seealso>
      can do this:</p>
    <pre>
6> <input>logger:add_handler_filter(myhandler, stop_non_info,
                             {fun logger_filters:level/2, {stop, neq, info}}).</input>
ok</pre>
    <p>See section <seealso marker="#filters">Filters</seealso> for
      more information about the filters and the <c>filter_default</c>
      configuration parameter.</p>

  </section>

  <section>
    <title>Example: Implement a handler</title>
    <p>Section <seealso marker="logger#handler_callback_functions">Handler
	Callback Functions</seealso> in the logger(3) manual page
	describes the callback functions that can be implemented for a
	Logger handler.</p>
    <p>A handler callback module must export:</p>
    <list>
      <item><c>log(Log, Config)</c></item>
    </list>
    <p>It can optionally also export some, or all, of the following:</p>
    <list>
      <item><c>adding_handler(Config)</c></item>
      <item><c>removing_handler(Config)</c></item>
      <item><c>changing_config(SetOrUpdate, OldConfig, NewConfig)</c></item>
      <item><c>filter_config(Config)</c></item>
    </list>
    <p>When a handler is added, by for example a call
      to <seealso marker="logger#add_handler-3">
	<c>logger:add_handler(Id, HModule, Config)</c></seealso>,
      Logger first calls <c>HModule:adding_handler(Config)</c>. If
      this function returns <c>{ok,Config1}</c>, Logger
      writes <c>Config1</c> to the configuration database, and
      the <c>logger:add_handler/3</c> call returns. After this, the
      handler is installed and must be ready to receive log events as
      calls to <c>HModule:log/2</c>.</p>
    <p>A handler can be removed by calling
      <seealso marker="logger#remove_handler-1">
	<c>logger:remove_handler(Id)</c></seealso>. Logger calls
      <c>HModule:removing_handler(Config)</c>, and removes the
      handler's configuration from the configuration database.</p>
    <p>When <seealso marker="logger#set_handler_config-2">
	<c>logger:set_handler_config/2,3</c></seealso>
      or <seealso marker="logger#update_handler_config/2">
	<c>logger:update_handler_config/2,3</c></seealso> is called,
      Logger
      calls <c>HModule:changing_config(SetOrUpdate, OldConfig, NewConfig)</c>. If
      this function returns <c>{ok,NewConfig1}</c>, Logger
      writes <c>NewConfig1</c> to the configuration database.</p>
    <p>When <seealso marker="logger#get_config-0">
	<c>logger:get_config/0</c></seealso> or
      <seealso marker="logger#get_handler_config-0">
	<c>logger:get_handler_config/0,1</c></seealso> is called,
      Logger calls <c>HModule:filter_config(Config)</c>. This function
      must return the handler configuration where internal data is
      removed.</p>

    <p>A simple handler that prints to the terminal can be implemented
      as follows:</p>
    <code>
-module(myhandler1).
-export([log/2]).

log(LogEvent, #{formatter := {FModule, FConfig}}) ->
    io:put_chars(FModule:format(LogEvent, FConfig)).
    </code>

    <p>Notice that the above handler does not have any overload
      protection, and all log events are printed directly from the
      client process.</p>
    <p>For information and examples of overload protection, please
      refer to
      section <seealso marker="#overload_protection">Protecting the
	Handler from Overload</seealso>, and the implementation
      of <seealso marker="logger_std_h"><c>logger_std_h</c></seealso>
      and <seealso marker="logger_disk_log_h"><c>logger_disk_log_h</c>
      </seealso>.</p>
    <p>The following is a simpler example of a handler which logs to a
      file through one single process:</p>
    <code>
-module(myhandler2).
-export([adding_handler/1, removing_handler/1, log/2]).
-export([init/1, handle_call/3, handle_cast/2, terminate/2]).

adding_handler(Config) ->
    MyConfig = maps:get(config,Config,#{file => "myhandler2.log"}),
    {ok, Pid} = gen_server:start(?MODULE, MyConfig, []),
    {ok, Config#{config => MyConfig#{pid => Pid}}}.

removing_handler(#{config := #{pid := Pid}}) ->
    gen_server:stop(Pid).

log(LogEvent,#{config := #{pid := Pid}} = Config) ->
    gen_server:cast(Pid, {log, LogEvent, Config}).

init(#{file := File}) ->
    {ok, Fd} = file:open(File, [append, {encoding, utf8}]),
    {ok, #{file => File, fd => Fd}}.

handle_call(_, _, State) ->
    {reply, {error, bad_request}, State}.

handle_cast({log, LogEvent, Config}, #{fd := Fd} = State) ->
    do_log(Fd, LogEvent, Config),
    {noreply, State}.

terminate(_Reason, #{fd := Fd}) ->
    _ = file:close(Fd),
    ok.

do_log(Fd, LogEvent, #{formatter := {FModule, FConfig}}) ->
    String = FModule:format(LogEvent, FConfig),
    io:put_chars(Fd, String).
    </code>
  </section>

  <section>
    <marker id="overload_protection"/>
    <title>Protecting the Handler from Overload</title>
    <p>The default handlers, <seealso marker="logger_std_h">
    <c>logger_std_h</c></seealso> and <seealso marker="logger_disk_log_h">
    <c>logger_disk_log_h</c></seealso>, feature an overload protection
    mechanism, which makes it possible for the handlers to survive,
    and stay responsive, during periods of high load (when huge
    numbers of incoming log requests must be handled).
    The mechanism works as follows:</p>
    
    <section>
      <title>Message Queue Length</title>
      <p>The handler process keeps track of the length of its message
      queue and takes some form of action when the current length exceeds a
      configurable threshold. The purpose is to keep the handler in, or to
      as quickly as possible get the handler into, a state where it can
      keep up with the pace of incoming log events. The memory use of the
      handler must never grow larger and larger, since that will eventually
      cause the handler to crash. These three thresholds, with associated
      actions, exist:</p>
      
      <taglist>
	<tag><c>sync_mode_qlen</c></tag>
	<item>
	  <p>As long as the length of the message queue is lower than this
	  value, all log events are handled asynchronously. This means that
	  the client process sending the log event, by calling a log function
	  in the <seealso marker="logger_chapter#logger_api">Logger API</seealso>,
	  does not wait for a response from the handler but continues
	  executing immediately after the event is sent. It is not affected
	  by the time it takes the handler to print the event to the log
	  device. If the message queue grows larger than this value,
	  the handler starts handling log events synchronously instead,
	  meaning that the client process sending the event must wait for a
	  response. When the handler reduces the message queue to a
	  level below the <c>sync_mode_qlen</c> threshold, asynchronous
	  operation is resumed. The switch from asynchronous to synchronous
	  mode can slow down the logging tempo of one, or a few, busy senders,
	  but cannot protect the handler sufficiently in a situation of many
	  busy concurrent senders.</p>
	  <p>Defaults to <c>10</c> messages.</p>
	</item>
	<tag><c>drop_mode_qlen</c></tag>
	<item>
	  <p>When the message queue grows larger than this threshold, the
	  handler switches to a mode in which it drops all new events that
	  senders want to log. Dropping an event in this mode means that the
	  call to the log function never results in a message being sent to
	  the handler, but the function returns without taking any action.
	  The handler keeps logging the events that are already in its message
	  queue, and when the length of the message queue is reduced to a level
	  below the threshold, synchronous or asynchronous mode is resumed.
	  Notice that when the handler activates or deactivates drop mode,
	  information about it is printed in the log.</p>
	  <p>Defaults to <c>200</c> messages.</p>
	</item>
	<tag><c>flush_qlen</c></tag>
	<item>
	  <p>If the length of the message queue grows larger than this threshold,
	  a flush (delete) operation takes place. To flush events, the handler
	  discards the messages in the message queue by receiving them in a
	  loop without logging. Client processes waiting for a response from a
	  synchronous log request receive a reply from the handler indicating
	  that the request is dropped. The handler process increases its
	  priority during the flush loop to make sure that no new events
	  are received during the operation. Notice that after the flush operation
	  is performed, the handler prints information in the log about how many
	  events have been deleted.</p>
	  <p>Defaults to <c>1000</c> messages.</p>
	</item>
      </taglist>

      <p>For the overload protection algorithm to work properly, it is
      required that:</p>

      <p><c>sync_mode_qlen =&lt; drop_mode_qlen =&lt; flush_qlen</c></p>

      <p>and that:</p>

      <p><c>drop_mode_qlen &gt; 1</c></p>

      <p>To disable certain modes, do the following:</p>
      <list>
	<item>If <c>sync_mode_qlen</c> is set to <c>0</c>, all log events are handled
	synchronously. That is, asynchronous logging is disabled.</item>
	<item>If <c>sync_mode_qlen</c> is set to the same value as
	<c>drop_mode_qlen</c>, synchronous mode is disabled. That is, the handler
	always runs in asynchronous mode, unless dropping or flushing is invoked.</item>
	<item>If <c>drop_mode_qlen</c> is set to the same value as <c>flush_qlen</c>,
	drop mode is disabled and can never occur.</item>
      </list>

      <p>During high load scenarios, the length of the handler message queue
      rarely grows in a linear and predictable way. Instead, whenever the
      handler process is scheduled in, it can have an almost arbitrary number
      of messages waiting in the message queue. It is for this reason that the overload
      protection mechanism is focused on acting quickly, and quite drastically,
      such as immediately dropping or flushing messages, when a large queue length
      is detected.</p>

      <p>The values of the previously listed thresholds can be specified by the user.
      This way, a handler can be configured to, for example, not drop or flush
      messages unless the message queue length of the handler process grows extremely
      large. Notice that large amounts of memory can be required for the node under such
      circumstances. Another example of user configuration is when, for performance
      reasons, the client processes must never be blocked by synchronous log requests.
      It is possible, perhaps, that dropping or flushing events is still acceptable, since
      it does not affect the performance of the client processes sending the log events.</p>

      <p>A configuration example:</p>
      <code type="none">
logger:add_handler(my_standard_h, logger_std_h,
                   #{config => #{file => "./system_info.log",
                                 sync_mode_qlen => 100,
                                 drop_mode_qlen => 1000,
                                 flush_qlen => 2000}}).
    </code>
    </section>

    <section>
      <title>Controlling Bursts of Log Requests</title>
      <p>Large bursts of log events - many events received by the handler
      under a short period of time - can potentially cause problems, such as:</p>
      <list>
	<item>Log files grow very large, very quickly.</item>
	<item>Circular logs wrap too quickly so that important data is overwritten.</item>
	<item>Write buffers grow large, which slows down file sync operations.</item>
      </list>

      <p>For this reason, both built-in handlers offer the possibility to specify the
      maximum number of events to be handled within a certain time frame.
      With this burst control feature enabled, the handler can avoid choking the log with
      massive amounts of printouts. The configuration parameters are:</p>
      <taglist>
	<tag><c>burst_limit_enable</c></tag>
	<item>
	  <p>Value <c>true</c> enables burst control and <c>false</c> disables it.</p>
	  <p>Defaults to <c>true</c>.</p>
	</item>
	<tag><c>burst_limit_max_count</c></tag>
	<item>
	  <p>This is the maximum number of events to handle within a
	  <c>burst_limit_window_time</c> time frame. After the limit is
	  reached, successive events are dropped until the end of the time frame.</p>
	  <p>Defaults to <c>500</c> events.</p>
	</item>
	<tag><c>burst_limit_window_time</c></tag>
	<item>
	  <p>See the previous description of <c>burst_limit_max_count</c>.</p>
	  <p>Defaults to <c>1000</c> milliseconds.</p>
	</item>
      </taglist>

      <p>A configuration example:</p>
      <code type="none">
logger:add_handler(my_disk_log_h, logger_disk_log_h,
                   #{config => #{file => "./my_disk_log",
                                 burst_limit_enable => true,
                                 burst_limit_max_count => 20,
                                 burst_limit_window_time => 500}}).
    </code>
    </section>

    <section>
      <title>Terminating an Overloaded Handler</title>
      <p>It is possible that a handler, even if it can successfully manage peaks
      of high load without crashing, can build up a large message queue, or use a
      large amount of memory. The overload protection mechanism includes an
      automatic termination and restart feature for the purpose of guaranteeing
      that a handler does not grow out of bounds. The feature is configured
      with the following parameters:</p>
      <taglist>
	<tag><c>overload_kill_enable</c></tag>
	<item>
	  <p>Value <c>true</c> enables the feature and <c>false</c> disables it.</p>
	  <p>Defaults to <c>false</c>.</p>
	</item>
	<tag><c>overload_kill_qlen</c></tag>
	<item>
	  <p>This is the maximum allowed queue length. If the message queue grows
	  larger than this, the handler process is terminated.</p>
	  <p>Defaults to <c>20000</c> messages.</p>
	</item>
	<tag><c>overload_kill_mem_size</c></tag>
	<item>
	  <p>This is the maximum memory size that the handler process is allowed to use.
	  If the handler grows larger than this, the process is terminated.</p>
	  <p>Defaults to <c>3000000</c> bytes.</p>
	</item>
	<tag><c>overload_kill_restart_after</c></tag>
	<item>
	  <p>If the handler is terminated, it restarts automatically after a
	  delay specified in milliseconds. The value <c>infinity</c> prevents
	  restarts.</p>
	  <p>Defaults to <c>5000</c> milliseconds.</p>
	</item>
      </taglist>
      <p>If the handler process is terminated because of overload, it prints
      information about it in the log. It also prints information about when a
      restart has taken place, and the handler is back in action.</p>
      <note>
	<p>The sizes of the log events affect the memory needs of the handler.
	For information about how to limit the size of log events, see the
	<seealso marker="logger_formatter"><c>logger_formatter(3)</c></seealso>
	manual page.</p>
      </note>
    </section>
  </section>

  <section>
    <marker id="proxy"/>
    <title>Logger Proxy</title>
    <p>The Logger proxy is an Erlang process which is part of the
      Kernel application's supervision tree. During startup, the proxy
      process registers itself as the <c>system_logger</c>, meaning
      that log events produced by the emulator are sent to this
      process.</p>
    <p>When a log event is issued on a process which has its group
      leader on a remote node, Logger automatically forwards the log
      event to the group leader's node. To achieve this, it first
      sends the log event as an Erlang message from the original
      client process to the proxy on the local node, and the proxy in
      turn forwards the event to the proxy on the remote node.</p>
    <p>When receiving a log event, either from the emulator or from a
      remote node, the proxy calls the Logger API to log the event.</p>
    <p>The proxy process is overload protected in the same way as
      described in
      section <seealso marker="#overload_protection">Protecting the
      Handler from Overload</seealso>, but with the following default
      values:</p>
    <code>
    #{sync_mode_qlen => 500,
      drop_mode_qlen => 1000,
      flush_qlen => 5000,
      burst_limit_enable => false,
      overload_kill_enable => false}</code>
    <p>For log events from the emulator, synchronous message passing
      mode is not applicable, since all messages are passed
      asynchronously by the emulator. Drop mode is achieved by setting
      the <c>system_logger</c> to <c>undefined</c>, forcing the
      emulator to drop events until it is set back to the proxy pid
      again.</p>
    <p>The proxy uses <seealso marker="erts:erlang#send_nosuspend/2">
	<c>erlang:send_nosuspend/2</c></seealso> when sending log
      events to a remote node. If the message could not be sent
      without suspending the sender, it is dropped. This is to avoid
      blocking the proxy process.</p>
  </section>

  <section>
    <title>See Also</title>
    <p>
      <seealso marker="disk_log"><c>disk_log(3)</c></seealso>,
      <seealso marker="erts:erlang"><c>erlang(3)</c></seealso>,
      <seealso marker="error_logger"><c>error_logger(3)</c></seealso>,
      <seealso marker="logger"><c>logger(3)</c></seealso>,
      <seealso marker="logger_disk_log_h"><c>logger_disk_log_h(3)</c></seealso>,
      <seealso marker="logger_filters"><c>logger_filters(3)</c></seealso>,
      <seealso marker="logger_formatter"><c>logger_formatter(3)</c></seealso>,
      <seealso marker="logger_std_h"><c>logger_std_h(3)</c></seealso>,
        <seealso marker="sasl:sasl_app"><c>sasl(6)</c></seealso></p>
  </section>
</chapter>