Class: Fluent::KafkaOutputBuffered

Inherits:
BufferedOutput
  • Object
show all
Includes:
Fluent::KafkaPluginUtil::SSLSettings
Defined in:
lib/fluent/plugin/out_kafka_buffered.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Fluent::KafkaPluginUtil::SSLSettings

included, #read_ssl_file

Constructor Details

#initialize ⇒ KafkaOutputBuffered

Returns a new instance of KafkaOutputBuffered.



81
82
83
84
85
86
87
88
89
90
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 81

def initialize
  super

  require 'kafka'
  require 'fluent/plugin/kafka_producer_ext'

  @kafka = nil
  @producers = {}
  @producers_mutex = Mutex.new
end

Instance Attribute Details

#field_separator ⇒ Object

Returns the value of attribute field_separator.



75
76
77
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 75

def field_separator
  @field_separator
end

#output_data_type ⇒ Object

Returns the value of attribute output_data_type.



74
75
76
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 74

def output_data_type
  @output_data_type
end

Instance Method Details

#configure(conf) ⇒ Object 



121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 121

def configure(conf)
  super

  if @zookeeper
    require 'zookeeper'
  else
    @seed_brokers = @brokers.match(",").nil? ? [@brokers] : @brokers.split(",")
    log.info "brokers has been set directly: #{@seed_brokers}"
  end

  if conf['ack_timeout_ms']
    log.warn "'ack_timeout_ms' parameter is deprecated. Use second unit 'ack_timeout' instead"
    @ack_timeout = conf['ack_timeout_ms'].to_i / 1000
  end

  @f_separator = case @field_separator
                 when /SPACE/i then ' '
                 when /COMMA/i then ','
                 when /SOH/i then "\x01"
                 else "\t"
                 end

  @formatter_proc = setup_formatter(conf)

  @producer_opts = {max_retries: @max_send_retries, required_acks: @required_acks}
  @producer_opts[:ack_timeout] = @ack_timeout if @ack_timeout
  @producer_opts[:compression_codec] = @compression_codec.to_sym if @compression_codec
end

#emit(tag, es, chain) ⇒ Object 



161
162
163
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 161

def emit(tag, es, chain)
  super(tag, es, chain, tag)
end

#format_stream(tag, es) ⇒ Object 



165
166
167
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 165

def format_stream(tag, es)
  es.to_msgpack_stream
end

#get_producer ⇒ Object 



178
179
180
181
182
183
184
185
186
187
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 178

def get_producer
  @producers_mutex.synchronize {
    producer = @producers[Thread.current.object_id]
    unless producer
      producer = @kafka.producer(@producer_opts)
      @producers[Thread.current.object_id] = producer
    end
    producer
  }
end

#get_schema_from_redis_by_name(schema_name) ⇒ Object 



346
347
348
349
350
351
352
353
354
355
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 346

def get_schema_from_redis_by_name schema_name
  if stored_schema = $redis.get(schema_name)
    parsed_schema = JSON.parse($redis.get(schema_name))
    {
        'schema_id' => parsed_schema['schema_id'],
        'schema' => Avro::Schema.parse(parsed_schema['schema_json']),
        'field_types' => parsed_schema['field_types']
    }
  end
end

#init_redis ⇒ Object 



338
339
340
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 338

def init_redis
  $redis = Redis.new
end

#refresh_client(raise_error = true) ⇒ Object 



92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 92

def refresh_client(raise_error = true)
  if @zookeeper
    @seed_brokers = []
    z = Zookeeper.new(@zookeeper)
    z.get_children(:path => @zookeeper_path)[:children].each do |id|
      broker = Yajl.load(z.get(:path => @zookeeper_path + "/#{id}")[:data])
      @seed_brokers.push("#{broker['host']}:#{broker['port']}")
    end
    z.close
    log.info "brokers has been refreshed via Zookeeper: #{@seed_brokers}"
  end
  begin
    if @seed_brokers.length > 0
      logger = @get_kafka_client_log ? log : nil
      @kafka = Kafka.new(seed_brokers: @seed_brokers, client_id: @client_id, logger: logger, ssl_ca_cert: read_ssl_file(@ssl_ca_cert),
                         ssl_client_cert: read_ssl_file(@ssl_client_cert), ssl_client_cert_key: read_ssl_file(@ssl_client_cert_key))
      log.info "initialized kafka producer: #{@client_id}"
    else
      log.warn "No brokers found on Zookeeper"
    end
  rescue Exception => e
    if raise_error # During startup, error should be reported to engine and stop its phase for safety.
      raise e
    else
      log.error e
    end
  end
end

#set_schema_to_redis(schema_name, schema) ⇒ Object 



342
343
344
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 342

def set_schema_to_redis schema_name, schema
  $redis.set(schema_name, schema.to_json)
end

#setup_formatter(conf) ⇒ Object 



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
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 189

def setup_formatter(conf)
  if @output_data_type == 'json'
    require 'yajl'
    Proc.new { |tag, time, record| Yajl::Encoder.encode(record) }
  elsif @output_data_type == 'ltsv'
    require 'ltsv'
    Proc.new { |tag, time, record| LTSV.dump(record) }
  elsif @output_data_type == 'msgpack'
    require 'msgpack'
    Proc.new { |tag, time, record| record.to_msgpack }
  elsif @output_data_type == 'avro'
    require "avro_turf"
    require 'avro_turf/messaging'
    require "avro/builder"
    init_redis
    Proc.new do |tag, time, record|
      record = record.select{|key, value| !key.nil? && !key.empty?}.map do |k, v|
        [
            k.tr('[]-', '_').delete('$'),
            (v.is_a?(Fixnum) || v.is_a?(Float) || v.nil? ? v : v.to_s.force_encoding("UTF-8"))
        ]
      end.to_h
      timestamp = Time.new
      record['enchilada_time_with_format'] = timestamp.strftime("%Y-%m-%dT%H:%M:%S.%LZ")
      @topic_name = schema_name = "#{tag.to_s.tr('.$:', '_')}_#{Digest::MD5.new.hexdigest(record.keys.to_s)[0..5]}"

      avro = AvroTurf::Messaging.new(registry_url: @schema_registry)

      unless (stored_schema = get_schema_from_redis_by_name(schema_name))
        fields = record.map do |key, value|
          {
              'name' => key,
              'type' => ['null', (value.is_a?(Fixnum) ? 'int' : (value.is_a?(Float) ? 'float' : 'string'))]
          }
        end
        field_types = fields.map{|field| [field['name'], (field['type'] - ['null']).first]}.to_h
        fields << {"name" => "enchilada_timestamp", "type" => "long"}
        schema_json = {
            "type": "record",
            "name": schema_name,
            "fields": fields
        }.to_json
        registry = avro.instance_variable_get('@registry')
        schema = Avro::Schema.parse(schema_json)
        schema_id = registry.register("#{schema_name}-value", schema)

        stored_schema = {
            'schema_json' => schema_json,
            'schema_id' => schema_id,
            'field_types' => field_types,
            'schema' => schema
        }

        set_schema_to_redis(schema_name, stored_schema)

      end

      record['enchilada_timestamp'] = timestamp.strftime('%s%3N').to_i
      record = record.map do |key, val|
        [key, (stored_schema['field_types'][key] != 'string' || val.nil? ? val : val.to_s)]
      end.to_h

      avro.encode(record, stored_schema['schema_id'], schema: stored_schema['schema'])
    end
  elsif @output_data_type =~ /^attr:(.*)$/
    @custom_attributes = $1.split(',').map(&:strip).reject(&:empty?)
    @custom_attributes.unshift('time') if @output_include_time
    @custom_attributes.unshift('tag') if @output_include_tag
    Proc.new { |tag, time, record|
      @custom_attributes.map { |attr|
        record[attr].nil? ? '' : record[attr].to_s
      }.join(@f_separator)
    }
  else
    @formatter = Fluent::Plugin.new_formatter(@output_data_type)
    @formatter.configure(conf)
    @formatter.method(:format)
  end
end

#shutdown ⇒ Object 



155
156
157
158
159
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 155

def shutdown
  super
  shutdown_producers
  @kafka = nil
end

#shutdown_producers ⇒ Object 



169
170
171
172
173
174
175
176
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 169

def shutdown_producers
  @producers_mutex.synchronize {
    @producers.each { |key, producer|
      producer.shutdown
    }
    @producers = {}
  }
end

#start ⇒ Object 



150
151
152
153
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 150

def start
  super
  refresh_client
end

#write(chunk) ⇒ Object 



269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
 
# File 'lib/fluent/plugin/out_kafka_buffered.rb', line 269

def write(chunk)
  tag = chunk.key
  producer = get_producer

  records_by_topic = {}
  bytes_by_topic = {}
  messages = 0
  messages_bytes = 0
  record_buf = nil
  record_buf_bytes = nil

  begin
    chunk.msgpack_each { |time, record|
      begin
        if @output_include_time
          if @time_format
            record['time'.freeze] = Time.at(time).strftime(@time_format)
          else
            record['time'.freeze] = time
          end
        end

        record['tag'] = tag if @output_include_tag
        record_buf = @formatter_proc.call(tag, time, record)
        topic = (@exclude_topic_key ? record.delete('topic'.freeze) : record['topic'.freeze]) || @topic_name
        partition_key = (@exclude_partition_key ? record.delete('partition_key'.freeze) : record['partition_key'.freeze]) || @default_partition_key
        partition = (@exclude_partition ? record.delete('partition'.freeze) : record['partition'.freeze]) || @default_partition
        message_key = (@exclude_message_key ? record.delete('message_key'.freeze) : record['message_key'.freeze]) || @default_message_key

        records_by_topic[topic] ||= 0
        bytes_by_topic[topic] ||= 0

        record_buf = @formatter_proc.call(tag, time, record)
        record_buf_bytes = record_buf.bytesize
      rescue StandardError => e
        log.warn "unexpected error during format record. Skip broken event:", :error => e.to_s, :error_class => e.class.to_s, :time => time, :record => record
        next
      end

      if (messages > 0) and (messages_bytes + record_buf_bytes > @kafka_agg_max_bytes)
        log.on_trace { log.trace("#{messages} messages send.") }
        producer.deliver_messages
        messages = 0
        messages_bytes = 0
      end
      log.on_trace { log.trace("message will send to #{topic} with partition_key: #{partition_key}, partition: #{partition}, message_key: #{message_key} and value: #{record_buf}.") }
      messages += 1
      producer.produce2(record_buf, topic: topic, key: message_key, partition_key: partition_key, partition: partition)
      messages_bytes += record_buf_bytes

      records_by_topic[topic] += 1
      bytes_by_topic[topic] += record_buf_bytes
    }
    if messages > 0
      log.trace { "#{messages} messages send." }
      producer.deliver_messages
    end
    log.debug { "(records|bytes) (#{records_by_topic}|#{bytes_by_topic})" }
  end
rescue Exception => e
  log.warn "Send exception occurred: #{e}"
  log.warn "Exception Backtrace : #{e.backtrace.join("\n")}"
  # For safety, refresh client and its producers
  shutdown_producers
  refresh_client(false)
  # Raise exception to retry sendind messages
  raise e
end