Class: Fluent::Plugin::ElasticsearchOutput
- Inherits:
-
Output
- Object
- Output
- Fluent::Plugin::ElasticsearchOutput
- Includes:
- ElasticsearchIndexTemplate, ElasticsearchConstants
- Defined in:
- lib/fluent/plugin/out_elasticsearch.rb
Direct Known Subclasses
Defined Under Namespace
Classes: ConnectionFailure
Constant Summary collapse
- DEFAULT_BUFFER_TYPE =
"memory"
- DEFAULT_ELASTICSEARCH_VERSION =
For compatibility.
5
- DEFAULT_TYPE_NAME_ES_7x =
"_doc".freeze
- DEFAULT_TYPE_NAME =
"fluentd".freeze
Constants included from ElasticsearchConstants
Fluent::Plugin::ElasticsearchConstants::BODY_DELIMITER, Fluent::Plugin::ElasticsearchConstants::CREATE_OP, Fluent::Plugin::ElasticsearchConstants::ID_FIELD, Fluent::Plugin::ElasticsearchConstants::INDEX_OP, Fluent::Plugin::ElasticsearchConstants::TIMESTAMP_FIELD, Fluent::Plugin::ElasticsearchConstants::UPDATE_OP, Fluent::Plugin::ElasticsearchConstants::UPSERT_OP
Instance Method Summary collapse
- #append_record_to_messages(op, meta, header, record, msgs) ⇒ Object
- #client ⇒ Object
- #configure(conf) ⇒ Object
- #connection_options_description ⇒ Object
- #convert_compat_id_key(key) ⇒ Object
- #create_meta_config_map ⇒ Object
-
#create_time_parser ⇒ Object
once fluent v0.14 is released we might be able to use Fluent::Parser::TimeParser, but it doesn’t quite do what we want - if gives [sec,nsec] where as we want something we can call ‘strftime` on…
- #detect_es_major_version ⇒ Object
- #expand_placeholders(metadata) ⇒ Object
- #flatten_record(record, prefix = []) ⇒ Object
- #get_connection_options ⇒ Object
- #get_escaped_userinfo(host_str) ⇒ Object
-
#get_parent_of(record, path) ⇒ Object
returns [parent, child_key] of child described by path array in record’s tree returns [nil, child_key] if path doesnt exist in record.
-
#initialize ⇒ ElasticsearchOutput
constructor
A new instance of ElasticsearchOutput.
- #multi_workers_ready? ⇒ Boolean
- #parse_time(value, event_time, tag) ⇒ Object
- #process_message(tag, meta, header, time, record, bulk_message, extracted_values) ⇒ Object
- #remove_keys(record) ⇒ Object
- #send_bulk(data) ⇒ Object
- #update_body(record, op) ⇒ Object
- #write(chunk) ⇒ Object
Methods included from ElasticsearchIndexTemplate
#get_template, #template_exists?, #template_install, #template_put, #templates_hash_install
Constructor Details
#initialize ⇒ ElasticsearchOutput
Returns a new instance of ElasticsearchOutput.
98 99 100 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 98 def initialize super end |
Instance Method Details
#append_record_to_messages(op, meta, header, record, msgs) ⇒ Object
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 308 def (op, , header, record, msgs) case op when UPDATE_OP, UPSERT_OP if .has_key?(ID_FIELD) header[UPDATE_OP] = msgs << @dump_proc.call(header) << BODY_DELIMITER msgs << @dump_proc.call(update_body(record, op)) << BODY_DELIMITER end when CREATE_OP if .has_key?(ID_FIELD) header[CREATE_OP] = msgs << @dump_proc.call(header) << BODY_DELIMITER msgs << @dump_proc.call(record) << BODY_DELIMITER end when INDEX_OP header[INDEX_OP] = msgs << @dump_proc.call(header) << BODY_DELIMITER msgs << @dump_proc.call(record) << BODY_DELIMITER end end |
#client ⇒ Object
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 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 221 def client @_es ||= begin = { client_key: @client_key, client_cert: @client_cert, client_key_pass: @client_key_pass } adapter_conf = lambda {|f| f.adapter :excon, } transport = Elasticsearch::Transport::Transport::HTTP::Faraday.new(.merge( options: { reload_connections: @reload_connections, reload_on_failure: @reload_on_failure, resurrect_after: @resurrect_after, retry_on_failure: 5, logger: @transport_logger, transport_options: { headers: { 'Content-Type' => @content_type.to_s }, request: { timeout: @request_timeout }, ssl: { verify: @ssl_verify, ca_file: @ca_file, version: @ssl_version } }, http: { user: @user, password: @password } }), &adapter_conf) es = Elasticsearch::Client.new transport: transport begin raise ConnectionFailure, "Can not reach Elasticsearch cluster (#{connection_options_description})!" unless es.ping rescue *es.transport.host_unreachable_exceptions => e raise ConnectionFailure, "Can not reach Elasticsearch cluster (#{connection_options_description})! #{e.message}" end log.info "Connection opened to Elasticsearch cluster => #{connection_options_description}" es end end |
#configure(conf) ⇒ Object
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 102 def configure(conf) compat_parameters_convert(conf, :buffer) super raise Fluent::ConfigError, "'tag' in chunk_keys is required." if not @chunk_key_tag @time_parser = create_time_parser if @remove_keys @remove_keys = @remove_keys.split(/\s*,\s*/) end if @target_index_key && @target_index_key.is_a?(String) @target_index_key = @target_index_key.split '.' end if @target_type_key && @target_type_key.is_a?(String) @target_type_key = @target_type_key.split '.' end if @remove_keys_on_update && @remove_keys_on_update.is_a?(String) @remove_keys_on_update = @remove_keys_on_update.split ',' end if @template_name && @template_file template_install(@template_name, @template_file, @template_overwrite) elsif @templates templates_hash_install(@templates, @template_overwrite) end # Consider missing the prefix of "$." in nested key specifiers. @id_key = convert_compat_id_key(@id_key) if @id_key @parent_key = convert_compat_id_key(@parent_key) if @parent_key @routing_key = convert_compat_id_key(@routing_key) if @routing_key = begin require 'oj' @dump_proc = Oj.method(:dump) rescue LoadError @dump_proc = Yajl.method(:dump) end if @user && m = @user.match(/%{(?<user>.*)}/) @user = URI.encode_www_form_component(m["user"]) end if @password && m = @password.match(/%{(?<password>.*)}/) @password = URI.encode_www_form_component(m["password"]) end if @hash_config raise Fluent::ConfigError, "@hash_config.hash_id_key and id_key must be equal." unless @hash_config.hash_id_key == @id_key end @transport_logger = nil if @with_transporter_log @transport_logger = log log_level = conf['@log_level'] || conf['log_level'] log.warn "Consider to specify log_level with @log_level." unless log_level end @last_seen_major_version = detect_es_major_version rescue DEFAULT_ELASTICSEARCH_VERSION if @last_seen_major_version == 6 && @type_name != DEFAULT_TYPE_NAME_ES_7x log.info "Detected ES 6.x: ES 7.x will only accept `_doc` in type_name." end if @last_seen_major_version >= 7 && @type_name != DEFAULT_TYPE_NAME_ES_7x log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`." @type_name = '_doc'.freeze end end |
#connection_options_description ⇒ Object
300 301 302 303 304 305 306 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 300 def [:hosts].map do |host_info| attributes = host_info.dup attributes[:password] = 'obfuscated' if attributes.has_key?(:password) attributes.inspect end.join(', ') end |
#convert_compat_id_key(key) ⇒ Object
178 179 180 181 182 183 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 178 def convert_compat_id_key(key) if key.include?('.') && !key.start_with?('$[') key = "$.#{key}" unless key.start_with?('$.') end key end |
#create_meta_config_map ⇒ Object
185 186 187 188 189 190 191 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 185 def result = [] result << [record_accessor_create(@id_key), '_id'] if @id_key result << [record_accessor_create(@parent_key), '_parent'] if @parent_key result << [record_accessor_create(@routing_key), '_routing'] if @routing_key result end |
#create_time_parser ⇒ Object
once fluent v0.14 is released we might be able to use Fluent::Parser::TimeParser, but it doesn’t quite do what we want - if gives
- sec,nsec
-
where as we want something we can call ‘strftime` on…
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 196 def create_time_parser if @time_key_format begin # Strptime doesn't support all formats, but for those it does it's # blazingly fast. strptime = Strptime.new(@time_key_format) Proc.new { |value| strptime.exec(value).to_datetime } rescue # Can happen if Strptime doesn't recognize the format; or # if strptime couldn't be required (because it's not installed -- it's # ruby 2 only) Proc.new { |value| DateTime.strptime(value, @time_key_format) } end else Proc.new { |value| DateTime.parse(value) } end end |
#detect_es_major_version ⇒ Object
173 174 175 176 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 173 def detect_es_major_version @_es_info ||= client.info @_es_info["version"]["number"].to_i end |
#expand_placeholders(metadata) ⇒ Object
366 367 368 369 370 371 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 366 def () logstash_prefix = extract_placeholders(@logstash_prefix, ) index_name = extract_placeholders(@index_name, ) type_name = extract_placeholders(@type_name, ) return logstash_prefix, index_name, type_name end |
#flatten_record(record, prefix = []) ⇒ Object
351 352 353 354 355 356 357 358 359 360 361 362 363 364 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 351 def flatten_record(record, prefix=[]) ret = {} if record.is_a? Hash record.each { |key, value| ret.merge! flatten_record(value, prefix + [key.to_s]) } elsif record.is_a? Array # Don't mess with arrays, leave them unprocessed ret.merge!({prefix.join(@flatten_hashes_separator) => record}) else return {prefix.join(@flatten_hashes_separator) => record} end ret end |
#get_connection_options ⇒ Object
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 267 def raise "`password` must be present if `user` is present" if @user && !@password hosts = if @hosts @hosts.split(',').map do |host_str| # Support legacy hosts format host:port,host:port,host:port... if host_str.match(%r{^[^:]+(\:\d+)?$}) { host: host_str.split(':')[0], port: (host_str.split(':')[1] || @port).to_i, scheme: @scheme } else # New hosts format expects URLs such as http://logs.foo.com,https://john:[email protected]/elastic uri = URI(get_escaped_userinfo(host_str)) %w(user password path).inject(host: uri.host, port: uri.port, scheme: uri.scheme) do |hash, key| hash[key.to_sym] = uri.public_send(key) unless uri.public_send(key).nil? || uri.public_send(key) == '' hash end end end.compact else [{host: @host, port: @port, scheme: @scheme}] end.each do |host| host.merge!(user: @user, password: @password) if !host[:user] && @user host.merge!(path: @path) if !host[:path] && @path end { hosts: hosts } end |
#get_escaped_userinfo(host_str) ⇒ Object
255 256 257 258 259 260 261 262 263 264 265 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 255 def get_escaped_userinfo(host_str) if m = host_str.match(/(?<scheme>.*)%{(?<user>.*)}:%{(?<password>.*)}(?<path>@.*)/) m["scheme"] + URI.encode_www_form_component(m["user"]) + ':' + URI.encode_www_form_component(m["password"]) + m["path"] else host_str end end |
#get_parent_of(record, path) ⇒ Object
returns [parent, child_key] of child described by path array in record’s tree returns [nil, child_key] if path doesnt exist in record
487 488 489 490 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 487 def get_parent_of(record, path) parent_object = path[0..-2].reduce(record) { |a, e| a.is_a?(Hash) ? a[e] : nil } [parent_object, path[-1]] end |
#multi_workers_ready? ⇒ Boolean
373 374 375 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 373 def multi_workers_ready? true end |
#parse_time(value, event_time, tag) ⇒ Object
214 215 216 217 218 219 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 214 def parse_time(value, event_time, tag) @time_parser.call(value) rescue => e router.emit_error_event(@time_parse_error_tag, Fluent::Engine.now, {'tag' => tag, 'time' => event_time, 'format' => @time_key_format, 'value' => value}, e) return Time.at(event_time).to_datetime end |
#process_message(tag, meta, header, time, record, bulk_message, extracted_values) ⇒ Object
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 401 def (tag, , header, time, record, , extracted_values) logstash_prefix, index_name, type_name = extracted_values if @flatten_hashes record = flatten_record(record) end if @hash_config record = generate_hash_id_key(record) end dt = nil if @logstash_format || if record.has_key?(TIMESTAMP_FIELD) rts = record[TIMESTAMP_FIELD] dt = parse_time(rts, time, tag) elsif record.has_key?(@time_key) rts = record[@time_key] dt = parse_time(rts, time, tag) record[TIMESTAMP_FIELD] = dt.iso8601(@time_precision) unless else dt = Time.at(time).to_datetime record[TIMESTAMP_FIELD] = dt.iso8601(@time_precision) end end target_index_parent, target_index_child_key = @target_index_key ? get_parent_of(record, @target_index_key) : nil if target_index_parent && target_index_parent[target_index_child_key] target_index = target_index_parent.delete(target_index_child_key) elsif @logstash_format dt = dt.new_offset(0) if @utc_index target_index = "#{logstash_prefix}#{@logstash_prefix_separator}#{dt.strftime(@logstash_dateformat)}" else target_index = index_name end # Change target_index to lower-case since Elasticsearch doesn't # allow upper-case characters in index names. target_index = target_index.downcase if @include_tag_key record[@tag_key] = tag end target_type_parent, target_type_child_key = @target_type_key ? get_parent_of(record, @target_type_key) : nil if target_type_parent && target_type_parent[target_type_child_key] target_type = target_type_parent.delete(target_type_child_key) if @last_seen_major_version == 6 log.warn "Detected ES 6.x: `@type_name` will be used as the document `_type`." target_type = type_name elsif @last_seen_major_version >= 7 log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`." target_type = '_doc'.freeze end else if @last_seen_major_version >= 7 && target_type != DEFAULT_TYPE_NAME_ES_7x log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`." target_type = '_doc'.freeze else target_type = type_name end end .clear ["_index".freeze] = target_index ["_type".freeze] = target_type if @pipeline ["pipeline".freeze] = @pipeline end .each do |record_accessor, | if raw_value = record_accessor.call(record) [] = raw_value end end if @remove_keys @remove_keys.each { |key| record.delete(key) } end (@write_operation, , header, record, ) @error. += 1 end |
#remove_keys(record) ⇒ Object
342 343 344 345 346 347 348 349 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 342 def remove_keys(record) keys = record[@remove_keys_on_update_key] || @remove_keys_on_update || [] record.delete(@remove_keys_on_update_key) return record unless keys.any? record = record.dup keys.each { |key| record.delete(key) } record end |
#send_bulk(data) ⇒ Object
492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 492 def send_bulk(data) retries = 0 begin response = client.bulk body: data @error.handle_error(response) if response['errors'] rescue *client.transport.host_unreachable_exceptions => e if retries < 2 retries += 1 @_es = nil @_es_info = nil log.warn "Could not push logs to Elasticsearch, resetting connection and trying again. #{e.message}" sleep 2**retries retry end raise ConnectionFailure, "Could not push logs to Elasticsearch after #{retries} retries. #{e.message}" rescue Exception @_es = nil if @reconnect_on_error @_es_info = nil if @reconnect_on_error raise end end |
#update_body(record, op) ⇒ Object
329 330 331 332 333 334 335 336 337 338 339 340 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 329 def update_body(record, op) update = remove_keys(record) body = {"doc".freeze => update} if op == UPSERT_OP if update == record body["doc_as_upsert".freeze] = true else body[UPSERT_OP] = record end end body end |
#write(chunk) ⇒ Object
377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 |
# File 'lib/fluent/plugin/out_elasticsearch.rb', line 377 def write(chunk) = '' header = {} = {} tag = chunk..tag extracted_values = (chunk.) @error = Fluent::Plugin::ElasticsearchErrorHandler.new(self) @last_seen_major_version = detect_es_major_version rescue DEFAULT_ELASTICSEARCH_VERSION chunk.msgpack_each do |time, record| @error.records += 1 next unless record.is_a? Hash begin (tag, , header, time, record, , extracted_values) rescue => e router.emit_error_event(tag, time, record, e) end end send_bulk() unless .empty? .clear end |