Class: Fluent::ViaqDataModelFilter

Inherits:
Filter
  • Object
show all
Includes:
ViaqDataModelFilterSystemd
Defined in:
lib/fluent/plugin/filter_viaq_data_model.rb

Constant Summary collapse

DOT_REPLACE_CHAR_UNUSED = 
'UNUSED'
NUM_FIELDS_UNLIMITED = 
-1
NORMAL_LEVELS =

github.com/ViaQ/elasticsearch-templates/blob/master/namespaces/default.yml#L63

{
  'emerg'    => 'emerg',
  'panic'    => 'emerg',
  'alert'    => 'alert',
  'crit'     => 'crit',
  'critical' => 'crit',
  'err'      => 'err',
  'error'    => 'err',
  'warning'  => 'warning',
  'warn'     => 'warning',
  'notice'   => 'notice',
  'info'     => 'info',
  'debug'    => 'debug',
  'trace'    => 'trace',
  'unknown'  => 'unknown',
}
PRIORITY_LEVELS =

numeric levels for the PRIORITY field

{
  0 => 'emerg',
  1 => 'alert',
  2 => 'crit',
  3 => 'err',
  4 => 'warning',
  5 => 'notice',
  6 => 'info',
  7 => 'debug',
  8 => 'trace',
  9 => 'unknown',
  '0' => 'emerg',
  '1' => 'alert',
  '2' => 'crit',
  '3' => 'err',
  '4' => 'warning',
  '5' => 'notice',
  '6' => 'info',
  '7' => 'debug',
  '8' => 'trace',
  '9' => 'unknown',
}

Constants included from ViaqDataModelFilterSystemd

ViaqDataModelFilterSystemd::JOURNAL_FIELD_MAP_SYSTEMD_K, ViaqDataModelFilterSystemd::JOURNAL_FIELD_MAP_SYSTEMD_T, ViaqDataModelFilterSystemd::JOURNAL_FIELD_MAP_SYSTEMD_U, ViaqDataModelFilterSystemd::JOURNAL_TIME_FIELDS

Instance Method Summary collapse

Methods included from ViaqDataModelFilterSystemd

#process_journal_fields

Instance Method Details

#add_elasticsearch_index_name_field(tag, time, record) ⇒ Object 



381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 381

def add_elasticsearch_index_name_field(tag, time, record)
  found = false
  @elasticsearch_index_names.each do |ein|
    if ein.matcher.match(tag)
      found = true
      return unless ein.enabled
      if ein.name_type == :operations_full || ein.name_type == :project_full
        field_name = @elasticsearch_index_name_field
        need_time = true
      else
        field_name = @elasticsearch_index_prefix_field
        need_time = false
      end

      case ein.name_type
      when :operations_full, :operations_prefix
        prefix = ".operations"
      when :project_full, :project_prefix
        name, uuid = nil
        unless record['kubernetes'].nil?
          k8s = record['kubernetes']
          name = k8s['namespace_name']
          uuid = k8s['namespace_id']
          if name.nil?
            log.error("record cannot use elasticsearch index name type #{ein.name_type}: record is missing kubernetes.namespace_name field: #{record}")
          end
          if uuid.nil?
            log.error("record cannot use elasticsearch index name type #{ein.name_type}: record is missing kubernetes.namespace_id field: #{record}")
          end
        else
          log.error("record cannot use elasticsearch index name type #{ein.name_type}: record is missing kubernetes field: #{record}")
        end
        if name.nil? || uuid.nil?
          name = @orphaned_namespace_name
        end
        prefix = name == @orphaned_namespace_name ? @orphaned_namespace_name : "project.#{name}.#{uuid}"
      end

      if ENV['CDM_DEBUG']
        unless tag == ENV['CDM_DEBUG_IGNORE_TAG']
          log.error("prefix #{prefix} need_time #{need_time} time #{record[@dest_time_name]}")
        end
      end

      if need_time
        ts = DateTime.parse(record[@dest_time_name])
        record[field_name] = prefix + "." + ts.strftime("%Y.%m.%d")
      else
        record[field_name] = prefix
      end
      if ENV['CDM_DEBUG']
        unless tag == ENV['CDM_DEBUG_IGNORE_TAG']
          log.error("record[#{field_name}] = #{record[field_name]}")
        end
      end

      break
    end
  end
  unless found
    if ENV['CDM_DEBUG']
      unless tag == ENV['CDM_DEBUG_IGNORE_TAG']
        log.error("no match for tag #{tag}")
      end
    end
  end
end

#add_pipeline_metadata(tag, time, record) ⇒ Object 



368
369
370
371
372
373
374
375
376
377
378
379
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 368

def add_pipeline_metadata (tag, time, record)
  record['pipeline_metadata'] = {} unless record.key?('pipeline_metadata')
  pipeline_type = @pipeline_type.to_s
  # this will catch the case where pipeline_type doesn't exist, or is not a Hash
  record['pipeline_metadata'][pipeline_type] = {} unless record['pipeline_metadata'][pipeline_type].respond_to?(:fetch)
  record['pipeline_metadata'][pipeline_type]['ipaddr4'] = @ipaddr4
  record['pipeline_metadata'][pipeline_type]['ipaddr6'] = @ipaddr6
  record['pipeline_metadata'][pipeline_type]['inputname'] = 'fluent-plugin-systemd'
  record['pipeline_metadata'][pipeline_type]['name'] = 'fluentd'
  record['pipeline_metadata'][pipeline_type]['received_at'] = Time.now.utc.to_datetime.rfc3339(6)
  record['pipeline_metadata'][pipeline_type]['version'] = @pipeline_version
end

#check_for_match_and_format(tag, time, record) ⇒ Object 



344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 344

def check_for_match_and_format(tag, time, record)
  return unless @formatters
  return if @formatter_cache_nomatch[tag]
  fmtr = @formatter_cache[tag]
  unless fmtr
    idx = @formatters.index{|fmtr| fmtr.matcher.match(tag)}
    if idx && (fmtr = @formatters[idx]).enabled
      @formatter_cache[tag] = fmtr
    else
      @formatter_cache_nomatch[tag] = true
      return
    end
  end
  fmtr.fmtr_func.call(tag, time, record, fmtr.fmtr_type)

  if record[@dest_time_name].nil? && record['time'].nil?
    record['time'] = Time.at(time).utc.to_datetime.rfc3339(6)
  end

  if fmtr.fmtr_remove_keys
    fmtr.fmtr_remove_keys.each{|k| record.delete(k)}
  end
end

#configure(conf) ⇒ Object 



160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 160

def configure(conf)
  super
  @keep_fields = {}
  @default_keep_fields.each{|xx| @keep_fields[xx] = true}
  @extra_keep_fields.each{|xx| @keep_fields[xx] = true}
  @keep_empty_fields_hash = {}
  @keep_empty_fields.each do |xx|
    @keep_empty_fields_hash[xx] = true
    @keep_fields[xx] = true
  end
  if @use_undefined && @keep_fields.key?(@undefined_name)
    raise Fluent::ConfigError, "Do not put [#{@undefined_name}] in default_keep_fields or extra_keep_fields"
  end
  if (@rename_time || @rename_time_if_not_exist) && @use_undefined && !@keep_fields.key?(@src_time_name)
    raise Fluent::ConfigError, "Field [#{@src_time_name}] must be listed in default_keep_fields or extra_keep_fields"
  end
  @undefined_dot_replace_char = nil if @undefined_dot_replace_char == DOT_REPLACE_CHAR_UNUSED
  if @formatters
    @formatters.each do |fmtr|
      matcher = ViaqMatchClass.new(fmtr.tag, nil)
      fmtr.instance_eval{ @params[:matcher] = matcher }
      fmtr.instance_eval{ @params[:fmtr_type] = fmtr.type }
      if fmtr.remove_keys
        fmtr.instance_eval{ @params[:fmtr_remove_keys] = fmtr.remove_keys.split(',') }
      else
        fmtr.instance_eval{ @params[:fmtr_remove_keys] = nil }
      end
      case fmtr.type
      when :sys_journal, :k8s_journal
        fmtr_func = method(:process_journal_fields)
      when :sys_var_log
        fmtr_func = method(:process_sys_var_log_fields)
      when :k8s_json_file
        fmtr_func = method(:process_k8s_json_file_fields)
      end
      fmtr.instance_eval{ @params[:fmtr_func] = fmtr_func }
    end
    @formatter_cache = {}
    @formatter_cache_nomatch = {}
  end
  begin
    @docker_hostname = File.open('/etc/docker-hostname') { |f| f.readline }.rstrip
  rescue
    @docker_hostname = nil
  end
  @ipaddr4 = ENV['IPADDR4'] || '127.0.0.1'
  @ipaddr6 = ENV['IPADDR6'] || '::1'
  @pipeline_version = (ENV['FLUENTD_VERSION'] || 'unknown fluentd version') + ' ' + (ENV['DATA_VERSION'] || 'unknown data version')
  # create the elasticsearch index name tag matchers
  unless @elasticsearch_index_names.empty?
    @elasticsearch_index_names.each do |ein|
      matcher = ViaqMatchClass.new(ein.tag, nil)
      ein.instance_eval{ @params[:matcher] = matcher }
    end
  end
end

#delempty(thing) ⇒ Object

recursively delete empty fields and empty lists/hashes from thing



232
233
234
235
236
237
238
239
240
241
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 232

def delempty(thing)
  if thing.respond_to?(:delete_if)
    if thing.kind_of? Hash
      thing.delete_if{|k,v| v.nil? || isempty(delempty(v)) || isempty(v)}
    else # assume single element iterable
      thing.delete_if{|elem| elem.nil? || isempty(delempty(elem)) || isempty(elem)}
    end
  end
  thing
end

#filter(tag, time, record) ⇒ Object 



499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 499

def filter(tag, time, record)
  if ENV['CDM_DEBUG']
    unless tag == ENV['CDM_DEBUG_IGNORE_TAG']
      log.error("input #{time} #{tag} #{record}")
    end
  end

  check_for_match_and_format(tag, time, record)
  add_pipeline_metadata(tag, time, record)
  handle_undefined_fields(tag, time, record)
  # remove the field from record if it is not in the list of fields to keep and
  # it is empty
  record.delete_if{|k,v| !@keep_empty_fields_hash.key?(k) && (v.nil? || isempty(delempty(v)) || isempty(v))}
  # probably shouldn't remove everything . . .
  log.warn("Empty record! tag [#{tag}] time [#{time}]") if record.empty?
  # rename the time field
  if (@rename_time || @rename_time_if_missing) && record.key?(@src_time_name)
    val = record.delete(@src_time_name)
    unless @rename_time_if_missing && record.key?(@dest_time_name)
      record[@dest_time_name] = val
    end
  end

  if !@elasticsearch_index_names.empty?
    add_elasticsearch_index_name_field(tag, time, record)
  elsif ENV['CDM_DEBUG']
    unless tag == ENV['CDM_DEBUG_IGNORE_TAG']
      log.error("not adding elasticsearch index name or prefix")
    end
  end
  if ENV['CDM_DEBUG']
    unless tag == ENV['CDM_DEBUG_IGNORE_TAG']
      log.error("output #{time} #{tag} #{record}")
    end
  end
  record
end

#handle_undefined_fields(tag, time, record) ⇒ Object 



465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 465

def handle_undefined_fields(tag, time, record)
  if @undefined_to_string || @use_undefined || @undefined_dot_replace_char || (@undefined_max_num_fields > NUM_FIELDS_UNLIMITED)
    # undefined contains all of the fields not in keep_fields
    undefined_keys = record.keys - @keep_fields.keys
    return if undefined_keys.empty?
    if @undefined_max_num_fields > NUM_FIELDS_UNLIMITED && undefined_keys.length > @undefined_max_num_fields
      undefined = {}
      undefined_keys.each{|k|undefined[k] = record.delete(k)}
      record[@undefined_name] = JSON.dump(undefined)
    else
      if @use_undefined
        record[@undefined_name] = {}
        modify_hsh = record[@undefined_name]
      else
        modify_hsh = record
      end
      undefined_keys.each do |k|
        origk = k
        if @use_undefined
          modify_hsh[k] = record.delete(k)
        end
        if @undefined_dot_replace_char && k.index('.')
          newk = k.gsub('.', @undefined_dot_replace_char)
          modify_hsh[newk] = modify_hsh.delete(k)
          k = newk
        end
        if @undefined_to_string && !modify_hsh[k].is_a?(String)
          modify_hsh[k] = JSON.dump(modify_hsh[k])
        end
      end
    end
  end
end

#isempty(thing) ⇒ Object

if thing doesn’t respond to empty? then assume it isn’t empty e.g. 0.respond_to?(:empty?) == false - the FixNum 0 is not empty



227
228
229
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 227

def isempty(thing)
  thing.respond_to?(:empty?) && thing.empty?
end

#normalize_level(level, newlevel, stream = nil, priority = nil) ⇒ Object 



283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 283

def normalize_level(level, newlevel, stream=nil, priority=nil)
  # if the record already has a level field, and it looks like one of our well
  # known values, convert it to the canonical normalized form - otherwise,
  # preserve the value in string format
  retlevel = nil
  if !level.nil?
    unless (retlevel = NORMAL_LEVELS[level]) ||
           (level.respond_to?(:downcase) && (retlevel = NORMAL_LEVELS[level.downcase]))
      retlevel = level.to_s # don't know what it is - just convert to string
    end
  elsif stream == 'stdout'
    retlevel = 'info'
  elsif stream == 'stderr'
    retlevel = 'err'
  elsif !priority.nil?
    retlevel = PRIORITY_LEVELS[priority]
  else
    retlevel = NORMAL_LEVELS[newlevel]
  end
  retlevel || 'unknown'
end

#process_k8s_json_file_fields(tag, time, record, fmtr_type = nil) ⇒ Object 



322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 322

def process_k8s_json_file_fields(tag, time, record, fmtr_type = nil)
  record['message'] = record['message'] || record['log']
  record['level'] = normalize_level(record['level'], nil, record['stream'])
  if record.key?('kubernetes') && record['kubernetes'].respond_to?(:fetch) && \
     (k8shost = record['kubernetes'].fetch('host', nil))
    record['hostname'] = k8shost
  elsif @docker_hostname
    record['hostname'] = @docker_hostname
  end
  if record[@dest_time_name].nil? # e.g. already has @timestamp
    unless record['time'].nil?
      # convert from string - parses a wide variety of formats
      rectime = Time.parse(record['time'])
    else
      # convert from time_t
      rectime = Time.at(time)
    end
    record['time'] = rectime.utc.to_datetime.rfc3339(6)
  end
  transform_eventrouter(tag, record)
end

#process_sys_var_log_fields(tag, time, record, fmtr_type = nil) ⇒ Object 



305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 305

def process_sys_var_log_fields(tag, time, record, fmtr_type = nil)
  record['systemd'] = {"t" => {"PID" => record['pid']}, "u" => {"SYSLOG_IDENTIFIER" => record['ident']}}
  if record[@dest_time_name].nil? # e.g. already has @timestamp
    # handle the case where the time reported in /var/log/messages is for a previous year
    timeobj = Time.at(time)
    if timeobj > Time.now
      timeobj = Time.new((timeobj.year - 1), timeobj.month, timeobj.day, timeobj.hour, timeobj.min, timeobj.sec, timeobj.utc_offset)
    end
    record['time'] = timeobj.utc.to_datetime.rfc3339(6)
  end
  if record['host'].eql?('localhost') && @docker_hostname
    record['hostname'] = @docker_hostname
  else
    record['hostname'] = record['host']
  end
end

#shutdown ⇒ Object 



221
222
223
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 221

def shutdown
  super
end

#start ⇒ Object 



217
218
219
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 217

def start
  super
end

#transform_eventrouter(tag, record) ⇒ Object 



449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
 
# File 'lib/fluent/plugin/filter_viaq_data_model.rb', line 449

def transform_eventrouter(tag, record)
  return unless @process_kubernetes_events
  if record.key?("event") && record["event"].respond_to?(:key?)
    if record.key?("verb")
      record["event"]["verb"] = record.delete("verb")
    end
    record["kubernetes"] = {} unless record.key?("kubernetes")
    record["kubernetes"]["event"] = record.delete("event")
    if record["kubernetes"]["event"].key?('message')
      ((record['pipeline_metadata'] ||= {})[@pipeline_type.to_s] ||= {})['original_raw_message'] = record['message']
    end
    record['message'] = record["kubernetes"]["event"].delete("message")
    record['time'] = record["kubernetes"]["event"]["metadata"].delete("creationTimestamp") 
  end
end