fluent-plugin-elasticsearch 3.5.3 → 3.5.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.coveralls.yml +2 -2
- data/.editorconfig +9 -9
- data/.gitignore +18 -18
- data/.travis.yml +18 -18
- data/Gemfile +9 -9
- data/History.md +390 -384
- data/ISSUE_TEMPLATE.md +27 -27
- data/LICENSE.txt +201 -201
- data/PULL_REQUEST_TEMPLATE.md +10 -10
- data/README.md +1515 -1398
- data/Rakefile +11 -11
- data/appveyor.yml +30 -30
- data/fluent-plugin-elasticsearch.gemspec +31 -31
- data/lib/fluent/log-ext.rb +38 -38
- data/lib/fluent/plugin/elasticsearch_constants.rb +13 -13
- data/lib/fluent/plugin/elasticsearch_error.rb +5 -5
- data/lib/fluent/plugin/elasticsearch_error_handler.rb +122 -122
- data/lib/fluent/plugin/elasticsearch_index_template.rb +112 -110
- data/lib/fluent/plugin/elasticsearch_simple_sniffer.rb +10 -10
- data/lib/fluent/plugin/filter_elasticsearch_genid.rb +25 -25
- data/lib/fluent/plugin/oj_serializer.rb +22 -22
- data/lib/fluent/plugin/out_elasticsearch.rb +777 -776
- data/lib/fluent/plugin/out_elasticsearch_dynamic.rb +262 -262
- data/test/helper.rb +24 -24
- data/test/plugin/test_alias_template.json +8 -8
- data/test/plugin/test_elasticsearch_error_handler.rb +503 -503
- data/test/plugin/test_filter_elasticsearch_genid.rb +44 -44
- data/test/plugin/test_out_elasticsearch.rb +2720 -2720
- data/test/plugin/test_out_elasticsearch_dynamic.rb +1001 -1001
- data/test/plugin/test_template.json +23 -23
- data/test/test_log-ext.rb +35 -35
- metadata +4 -3
@@ -1,110 +1,112 @@
|
|
1
|
-
require 'fluent/error'
|
2
|
-
require_relative './elasticsearch_error'
|
3
|
-
|
4
|
-
module Fluent::ElasticsearchIndexTemplate
|
5
|
-
def get_template(template_file)
|
6
|
-
if !File.exists?(template_file)
|
7
|
-
raise "If you specify a template_name you must specify a valid template file (checked '#{template_file}')!"
|
8
|
-
end
|
9
|
-
file_contents = IO.read(template_file).gsub(/\n/,'')
|
10
|
-
JSON.parse(file_contents)
|
11
|
-
end
|
12
|
-
|
13
|
-
def get_custom_template(template_file, customize_template)
|
14
|
-
if !File.exists?(template_file)
|
15
|
-
raise "If you specify a template_name you must specify a valid template file (checked '#{template_file}')!"
|
16
|
-
end
|
17
|
-
file_contents = IO.read(template_file).gsub(/\n/,'')
|
18
|
-
customize_template.each do |key, value|
|
19
|
-
file_contents = file_contents.gsub(key,value.downcase)
|
20
|
-
end
|
21
|
-
JSON.parse(file_contents)
|
22
|
-
end
|
23
|
-
|
24
|
-
def template_exists?(name)
|
25
|
-
client.indices.get_template(:name => name)
|
26
|
-
return true
|
27
|
-
rescue Elasticsearch::Transport::Transport::Errors::NotFound
|
28
|
-
return false
|
29
|
-
end
|
30
|
-
|
31
|
-
def retry_operate(max_retries, fail_on_retry_exceed = true)
|
32
|
-
return unless block_given?
|
33
|
-
retries = 0
|
34
|
-
begin
|
35
|
-
yield
|
36
|
-
rescue *client.transport.host_unreachable_exceptions, Timeout::Error => e
|
37
|
-
@_es = nil
|
38
|
-
@_es_info = nil
|
39
|
-
if retries < max_retries
|
40
|
-
retries += 1
|
41
|
-
|
42
|
-
|
43
|
-
|
44
|
-
|
45
|
-
|
46
|
-
|
47
|
-
|
48
|
-
|
49
|
-
|
50
|
-
|
51
|
-
|
52
|
-
|
53
|
-
|
54
|
-
|
55
|
-
|
56
|
-
|
57
|
-
|
58
|
-
|
59
|
-
|
60
|
-
|
61
|
-
|
62
|
-
|
63
|
-
|
64
|
-
|
65
|
-
|
66
|
-
|
67
|
-
|
68
|
-
|
69
|
-
|
70
|
-
|
71
|
-
|
72
|
-
log.info("Template configured
|
73
|
-
|
74
|
-
|
75
|
-
|
76
|
-
|
77
|
-
|
78
|
-
|
79
|
-
|
80
|
-
|
81
|
-
|
82
|
-
|
83
|
-
|
84
|
-
|
85
|
-
|
86
|
-
log.info("Template configured
|
87
|
-
|
88
|
-
|
89
|
-
|
90
|
-
|
91
|
-
|
92
|
-
|
93
|
-
|
94
|
-
|
95
|
-
|
96
|
-
|
97
|
-
log.info("The alias '#{deflector_alias_name}' is
|
98
|
-
|
99
|
-
|
100
|
-
|
101
|
-
|
102
|
-
|
103
|
-
|
104
|
-
|
105
|
-
|
106
|
-
|
107
|
-
|
108
|
-
|
109
|
-
|
110
|
-
end
|
1
|
+
require 'fluent/error'
|
2
|
+
require_relative './elasticsearch_error'
|
3
|
+
|
4
|
+
module Fluent::ElasticsearchIndexTemplate
|
5
|
+
def get_template(template_file)
|
6
|
+
if !File.exists?(template_file)
|
7
|
+
raise "If you specify a template_name you must specify a valid template file (checked '#{template_file}')!"
|
8
|
+
end
|
9
|
+
file_contents = IO.read(template_file).gsub(/\n/,'')
|
10
|
+
JSON.parse(file_contents)
|
11
|
+
end
|
12
|
+
|
13
|
+
def get_custom_template(template_file, customize_template)
|
14
|
+
if !File.exists?(template_file)
|
15
|
+
raise "If you specify a template_name you must specify a valid template file (checked '#{template_file}')!"
|
16
|
+
end
|
17
|
+
file_contents = IO.read(template_file).gsub(/\n/,'')
|
18
|
+
customize_template.each do |key, value|
|
19
|
+
file_contents = file_contents.gsub(key,value.downcase)
|
20
|
+
end
|
21
|
+
JSON.parse(file_contents)
|
22
|
+
end
|
23
|
+
|
24
|
+
def template_exists?(name)
|
25
|
+
client.indices.get_template(:name => name)
|
26
|
+
return true
|
27
|
+
rescue Elasticsearch::Transport::Transport::Errors::NotFound
|
28
|
+
return false
|
29
|
+
end
|
30
|
+
|
31
|
+
def retry_operate(max_retries, fail_on_retry_exceed = true)
|
32
|
+
return unless block_given?
|
33
|
+
retries = 0
|
34
|
+
begin
|
35
|
+
yield
|
36
|
+
rescue *client.transport.host_unreachable_exceptions, Timeout::Error => e
|
37
|
+
@_es = nil
|
38
|
+
@_es_info = nil
|
39
|
+
if retries < max_retries
|
40
|
+
retries += 1
|
41
|
+
wait_seconds = 2**retries
|
42
|
+
sleep wait_seconds
|
43
|
+
log.warn "Could not communicate to Elasticsearch, resetting connection and trying again. #{e.message}"
|
44
|
+
log.warn "Remaining retry: #{max_retries - retries}. Retry to communicate after #{wait_seconds} second(s)."
|
45
|
+
retry
|
46
|
+
end
|
47
|
+
message = "Could not communicate to Elasticsearch after #{retries} retries. #{e.message}"
|
48
|
+
log.warn message
|
49
|
+
raise Fluent::Plugin::ElasticsearchError::RetryableOperationExhaustedFailure,
|
50
|
+
message if fail_on_retry_exceed
|
51
|
+
end
|
52
|
+
end
|
53
|
+
|
54
|
+
def template_put(name, template)
|
55
|
+
client.indices.put_template(:name => name, :body => template)
|
56
|
+
end
|
57
|
+
|
58
|
+
def indexcreation(index_name)
|
59
|
+
client.indices.create(:index => index_name)
|
60
|
+
rescue Elasticsearch::Transport::Transport::Error => e
|
61
|
+
log.error("Error while index creation - #{index_name}: #{e.inspect}")
|
62
|
+
end
|
63
|
+
|
64
|
+
def template_install(name, template_file, overwrite)
|
65
|
+
if overwrite
|
66
|
+
template_put(name, get_template(template_file))
|
67
|
+
log.info("Template '#{name}' overwritten with #{template_file}.")
|
68
|
+
return
|
69
|
+
end
|
70
|
+
if !template_exists?(name)
|
71
|
+
template_put(name, get_template(template_file))
|
72
|
+
log.info("Template configured, but no template installed. Installed '#{name}' from #{template_file}.")
|
73
|
+
else
|
74
|
+
log.info("Template configured and already installed.")
|
75
|
+
end
|
76
|
+
end
|
77
|
+
|
78
|
+
def template_custom_install(template_name, template_file, overwrite, customize_template, index_prefix, rollover_index, deflector_alias_name, app_name, index_date_pattern)
|
79
|
+
template_custom_name=template_name.downcase
|
80
|
+
if overwrite
|
81
|
+
template_put(template_custom_name, get_custom_template(template_file, customize_template))
|
82
|
+
log.info("Template '#{template_custom_name}' overwritten with #{template_file}.")
|
83
|
+
else
|
84
|
+
if !template_exists?(template_custom_name)
|
85
|
+
template_put(template_custom_name, get_custom_template(template_file, customize_template))
|
86
|
+
log.info("Template configured, but no template installed. Installed '#{template_custom_name}' from #{template_file}.")
|
87
|
+
else
|
88
|
+
log.info("Template configured and already installed.")
|
89
|
+
end
|
90
|
+
end
|
91
|
+
|
92
|
+
if rollover_index
|
93
|
+
if !client.indices.exists_alias(:name => deflector_alias_name)
|
94
|
+
index_name_temp='<'+index_prefix.downcase+'-'+app_name.downcase+'-{'+index_date_pattern+'}-000001>'
|
95
|
+
indexcreation(index_name_temp)
|
96
|
+
client.indices.put_alias(:index => index_name_temp, :name => deflector_alias_name)
|
97
|
+
log.info("The alias '#{deflector_alias_name}' is created for the index '#{index_name_temp}'")
|
98
|
+
else
|
99
|
+
log.info("The alias '#{deflector_alias_name}' is already present")
|
100
|
+
end
|
101
|
+
else
|
102
|
+
log.info("No index and alias creation action performed because rollover_index is set to '#{rollover_index}'")
|
103
|
+
end
|
104
|
+
end
|
105
|
+
|
106
|
+
def templates_hash_install(templates, overwrite)
|
107
|
+
templates.each do |key, value|
|
108
|
+
template_install(key, value, overwrite)
|
109
|
+
end
|
110
|
+
end
|
111
|
+
|
112
|
+
end
|
@@ -1,10 +1,10 @@
|
|
1
|
-
require 'elasticsearch'
|
2
|
-
|
3
|
-
class Fluent::Plugin::ElasticsearchSimpleSniffer < Elasticsearch::Transport::Transport::Sniffer
|
4
|
-
|
5
|
-
def hosts
|
6
|
-
@transport.logger.debug "In Fluent::Plugin::ElasticsearchSimpleSniffer hosts #{@transport.hosts}" if @transport.logger
|
7
|
-
@transport.hosts
|
8
|
-
end
|
9
|
-
|
10
|
-
end
|
1
|
+
require 'elasticsearch'
|
2
|
+
|
3
|
+
class Fluent::Plugin::ElasticsearchSimpleSniffer < Elasticsearch::Transport::Transport::Sniffer
|
4
|
+
|
5
|
+
def hosts
|
6
|
+
@transport.logger.debug "In Fluent::Plugin::ElasticsearchSimpleSniffer hosts #{@transport.hosts}" if @transport.logger
|
7
|
+
@transport.hosts
|
8
|
+
end
|
9
|
+
|
10
|
+
end
|
@@ -1,25 +1,25 @@
|
|
1
|
-
require 'securerandom'
|
2
|
-
require 'base64'
|
3
|
-
require 'fluent/plugin/filter'
|
4
|
-
|
5
|
-
module Fluent::Plugin
|
6
|
-
class ElasticsearchGenidFilter < Filter
|
7
|
-
Fluent::Plugin.register_filter('elasticsearch_genid', self)
|
8
|
-
|
9
|
-
config_param :hash_id_key, :string, :default => '_hash'
|
10
|
-
|
11
|
-
def initialize
|
12
|
-
super
|
13
|
-
end
|
14
|
-
|
15
|
-
def configure(conf)
|
16
|
-
super
|
17
|
-
end
|
18
|
-
|
19
|
-
def filter(tag, time, record)
|
20
|
-
record[@hash_id_key] = Base64.strict_encode64(SecureRandom.uuid)
|
21
|
-
record
|
22
|
-
end
|
23
|
-
|
24
|
-
end
|
25
|
-
end
|
1
|
+
require 'securerandom'
|
2
|
+
require 'base64'
|
3
|
+
require 'fluent/plugin/filter'
|
4
|
+
|
5
|
+
module Fluent::Plugin
|
6
|
+
class ElasticsearchGenidFilter < Filter
|
7
|
+
Fluent::Plugin.register_filter('elasticsearch_genid', self)
|
8
|
+
|
9
|
+
config_param :hash_id_key, :string, :default => '_hash'
|
10
|
+
|
11
|
+
def initialize
|
12
|
+
super
|
13
|
+
end
|
14
|
+
|
15
|
+
def configure(conf)
|
16
|
+
super
|
17
|
+
end
|
18
|
+
|
19
|
+
def filter(tag, time, record)
|
20
|
+
record[@hash_id_key] = Base64.strict_encode64(SecureRandom.uuid)
|
21
|
+
record
|
22
|
+
end
|
23
|
+
|
24
|
+
end
|
25
|
+
end
|
@@ -1,22 +1,22 @@
|
|
1
|
-
require 'oj'
|
2
|
-
|
3
|
-
module Fluent::Plugin
|
4
|
-
module Serializer
|
5
|
-
|
6
|
-
class Oj
|
7
|
-
include Elasticsearch::Transport::Transport::Serializer::Base
|
8
|
-
|
9
|
-
# De-serialize a Hash from JSON string
|
10
|
-
#
|
11
|
-
def load(string, options={})
|
12
|
-
::Oj.load(string, options)
|
13
|
-
end
|
14
|
-
|
15
|
-
# Serialize a Hash to JSON string
|
16
|
-
#
|
17
|
-
def dump(object, options={})
|
18
|
-
::Oj.dump(object, options)
|
19
|
-
end
|
20
|
-
end
|
21
|
-
end
|
22
|
-
end
|
1
|
+
require 'oj'
|
2
|
+
|
3
|
+
module Fluent::Plugin
|
4
|
+
module Serializer
|
5
|
+
|
6
|
+
class Oj
|
7
|
+
include Elasticsearch::Transport::Transport::Serializer::Base
|
8
|
+
|
9
|
+
# De-serialize a Hash from JSON string
|
10
|
+
#
|
11
|
+
def load(string, options={})
|
12
|
+
::Oj.load(string, options)
|
13
|
+
end
|
14
|
+
|
15
|
+
# Serialize a Hash to JSON string
|
16
|
+
#
|
17
|
+
def dump(object, options={})
|
18
|
+
::Oj.dump(object, options)
|
19
|
+
end
|
20
|
+
end
|
21
|
+
end
|
22
|
+
end
|
@@ -1,776 +1,777 @@
|
|
1
|
-
# encoding: UTF-8
|
2
|
-
require 'date'
|
3
|
-
require 'excon'
|
4
|
-
require 'elasticsearch'
|
5
|
-
require 'json'
|
6
|
-
require 'uri'
|
7
|
-
begin
|
8
|
-
require 'strptime'
|
9
|
-
rescue LoadError
|
10
|
-
end
|
11
|
-
|
12
|
-
require 'fluent/plugin/output'
|
13
|
-
require 'fluent/event'
|
14
|
-
require 'fluent/error'
|
15
|
-
require 'fluent/time'
|
16
|
-
require 'fluent/log-ext'
|
17
|
-
require_relative 'elasticsearch_constants'
|
18
|
-
require_relative 'elasticsearch_error'
|
19
|
-
require_relative 'elasticsearch_error_handler'
|
20
|
-
require_relative 'elasticsearch_index_template'
|
21
|
-
begin
|
22
|
-
require_relative 'oj_serializer'
|
23
|
-
rescue LoadError
|
24
|
-
end
|
25
|
-
|
26
|
-
module Fluent::Plugin
|
27
|
-
class ElasticsearchOutput < Output
|
28
|
-
class RecoverableRequestFailure < StandardError; end
|
29
|
-
class UnrecoverableRequestFailure < Fluent::UnrecoverableError; end
|
30
|
-
|
31
|
-
# MissingIdFieldError is raised for records that do not
|
32
|
-
# include the field for the unique record identifier
|
33
|
-
class MissingIdFieldError < StandardError; end
|
34
|
-
|
35
|
-
# RetryStreamError privides a stream to be
|
36
|
-
# put back in the pipeline for cases where a bulk request
|
37
|
-
# failed (e.g some records succeed while others failed)
|
38
|
-
class RetryStreamError < StandardError
|
39
|
-
attr_reader :retry_stream
|
40
|
-
def initialize(retry_stream)
|
41
|
-
@retry_stream = retry_stream
|
42
|
-
end
|
43
|
-
end
|
44
|
-
|
45
|
-
RequestInfo = Struct.new(:host, :index)
|
46
|
-
|
47
|
-
helpers :event_emitter, :compat_parameters, :record_accessor
|
48
|
-
|
49
|
-
Fluent::Plugin.register_output('elasticsearch', self)
|
50
|
-
|
51
|
-
DEFAULT_BUFFER_TYPE = "memory"
|
52
|
-
DEFAULT_ELASTICSEARCH_VERSION = 5 # For compatibility.
|
53
|
-
DEFAULT_TYPE_NAME_ES_7x = "_doc".freeze
|
54
|
-
DEFAULT_TYPE_NAME = "fluentd".freeze
|
55
|
-
DEFAULT_RELOAD_AFTER = -1
|
56
|
-
TARGET_BULK_BYTES = 20 * 1024 * 1024
|
57
|
-
|
58
|
-
config_param :host, :string, :default => 'localhost'
|
59
|
-
config_param :port, :integer, :default => 9200
|
60
|
-
config_param :user, :string, :default => nil
|
61
|
-
config_param :password, :string, :default => nil, :secret => true
|
62
|
-
config_param :path, :string, :default => nil
|
63
|
-
config_param :scheme, :enum, :list => [:https, :http], :default => :http
|
64
|
-
config_param :hosts, :string, :default => nil
|
65
|
-
config_param :target_index_key, :string, :default => nil
|
66
|
-
config_param :target_type_key, :string, :default => nil,
|
67
|
-
:deprecated => <<EOC
|
68
|
-
Elasticsearch 7.x or above will ignore this config. Please use fixed type_name instead.
|
69
|
-
EOC
|
70
|
-
config_param :time_key_format, :string, :default => nil
|
71
|
-
config_param :time_precision, :integer, :default => 9
|
72
|
-
config_param :include_timestamp, :bool, :default => false
|
73
|
-
config_param :logstash_format, :bool, :default => false
|
74
|
-
config_param :logstash_prefix, :string, :default => "logstash"
|
75
|
-
config_param :logstash_prefix_separator, :string, :default => '-'
|
76
|
-
config_param :logstash_dateformat, :string, :default => "%Y.%m.%d"
|
77
|
-
config_param :utc_index, :bool, :default => true
|
78
|
-
config_param :type_name, :string, :default => DEFAULT_TYPE_NAME
|
79
|
-
config_param :index_name, :string, :default => "fluentd"
|
80
|
-
config_param :id_key, :string, :default => nil
|
81
|
-
config_param :write_operation, :string, :default => "index"
|
82
|
-
config_param :parent_key, :string, :default => nil
|
83
|
-
config_param :routing_key, :string, :default => nil
|
84
|
-
config_param :request_timeout, :time, :default => 5
|
85
|
-
config_param :reload_connections, :bool, :default => true
|
86
|
-
config_param :reload_on_failure, :bool, :default => false
|
87
|
-
config_param :retry_tag, :string, :default=>nil
|
88
|
-
config_param :resurrect_after, :time, :default => 60
|
89
|
-
config_param :time_key, :string, :default => nil
|
90
|
-
config_param :time_key_exclude_timestamp, :bool, :default => false
|
91
|
-
config_param :ssl_verify , :bool, :default => true
|
92
|
-
config_param :client_key, :string, :default => nil
|
93
|
-
config_param :client_cert, :string, :default => nil
|
94
|
-
config_param :client_key_pass, :string, :default => nil
|
95
|
-
config_param :ca_file, :string, :default => nil
|
96
|
-
config_param :ssl_version, :enum, list: [:SSLv23, :TLSv1, :TLSv1_1, :TLSv1_2], :default => :TLSv1
|
97
|
-
config_param :remove_keys, :string, :default => nil
|
98
|
-
config_param :remove_keys_on_update, :string, :default => ""
|
99
|
-
config_param :remove_keys_on_update_key, :string, :default => nil
|
100
|
-
config_param :flatten_hashes, :bool, :default => false
|
101
|
-
config_param :flatten_hashes_separator, :string, :default => "_"
|
102
|
-
config_param :template_name, :string, :default => nil
|
103
|
-
config_param :template_file, :string, :default => nil
|
104
|
-
config_param :template_overwrite, :bool, :default => false
|
105
|
-
config_param :customize_template, :hash, :default => nil
|
106
|
-
config_param :rollover_index, :string, :default => false
|
107
|
-
config_param :index_date_pattern, :string, :default => "now/d"
|
108
|
-
config_param :deflector_alias, :string, :default => nil
|
109
|
-
config_param :index_prefix, :string, :default => "logstash"
|
110
|
-
config_param :application_name, :string, :default => "default"
|
111
|
-
config_param :templates, :hash, :default => nil
|
112
|
-
config_param :max_retry_putting_template, :integer, :default => 10
|
113
|
-
config_param :fail_on_putting_template_retry_exceed, :bool, :default => true
|
114
|
-
config_param :max_retry_get_es_version, :integer, :default => 15
|
115
|
-
config_param :include_tag_key, :bool, :default => false
|
116
|
-
config_param :tag_key, :string, :default => 'tag'
|
117
|
-
config_param :time_parse_error_tag, :string, :default => 'Fluent::ElasticsearchOutput::TimeParser.error'
|
118
|
-
config_param :reconnect_on_error, :bool, :default => false
|
119
|
-
config_param :pipeline, :string, :default => nil
|
120
|
-
config_param :with_transporter_log, :bool, :default => false
|
121
|
-
config_param :emit_error_for_missing_id, :bool, :default => false
|
122
|
-
config_param :sniffer_class_name, :string, :default => nil
|
123
|
-
config_param :reload_after, :integer, :default => DEFAULT_RELOAD_AFTER
|
124
|
-
config_param :content_type, :enum, list: [:"application/json", :"application/x-ndjson"], :default => :"application/json",
|
125
|
-
:deprecated => <<EOC
|
126
|
-
elasticsearch gem v6.0.2 starts to use correct Content-Type. Please upgrade elasticserach gem and stop to use this option.
|
127
|
-
see: https://github.com/elastic/elasticsearch-ruby/pull/514
|
128
|
-
EOC
|
129
|
-
config_param :include_index_in_url, :bool, :default => false
|
130
|
-
config_param :http_backend, :enum, list: [:excon, :typhoeus], :default => :excon
|
131
|
-
config_param :validate_client_version, :bool, :default => false
|
132
|
-
config_param :prefer_oj_serializer, :bool, :default => false
|
133
|
-
config_param :unrecoverable_error_types, :array, :default => ["out_of_memory_error", "es_rejected_execution_exception"]
|
134
|
-
config_param :verify_es_version_at_startup, :bool, :default => true
|
135
|
-
config_param :default_elasticsearch_version, :integer, :default => DEFAULT_ELASTICSEARCH_VERSION
|
136
|
-
config_param :log_es_400_reason, :bool, :default => false
|
137
|
-
config_param :custom_headers, :hash, :default => {}
|
138
|
-
config_param :suppress_doc_wrap, :bool, :default => false
|
139
|
-
config_param :ignore_exceptions, :array, :default => [], value_type: :string, :desc => "Ignorable exception list"
|
140
|
-
config_param :exception_backup, :bool, :default => true, :desc => "Chunk backup flag when ignore exception occured"
|
141
|
-
config_param :bulk_message_request_threshold, :size, :default => TARGET_BULK_BYTES
|
142
|
-
|
143
|
-
config_section :buffer do
|
144
|
-
config_set_default :@type, DEFAULT_BUFFER_TYPE
|
145
|
-
config_set_default :chunk_keys, ['tag']
|
146
|
-
config_set_default :timekey_use_utc, true
|
147
|
-
end
|
148
|
-
|
149
|
-
include Fluent::ElasticsearchIndexTemplate
|
150
|
-
include Fluent::Plugin::ElasticsearchConstants
|
151
|
-
|
152
|
-
def initialize
|
153
|
-
super
|
154
|
-
end
|
155
|
-
|
156
|
-
def configure(conf)
|
157
|
-
compat_parameters_convert(conf, :buffer)
|
158
|
-
|
159
|
-
super
|
160
|
-
raise Fluent::ConfigError, "'tag' in chunk_keys is required." if not @chunk_key_tag
|
161
|
-
|
162
|
-
@time_parser = create_time_parser
|
163
|
-
@backend_options = backend_options
|
164
|
-
|
165
|
-
if @remove_keys
|
166
|
-
@remove_keys = @remove_keys.split(/\s*,\s*/)
|
167
|
-
end
|
168
|
-
|
169
|
-
if @target_index_key && @target_index_key.is_a?(String)
|
170
|
-
@target_index_key = @target_index_key.split '.'
|
171
|
-
end
|
172
|
-
|
173
|
-
if @target_type_key && @target_type_key.is_a?(String)
|
174
|
-
@target_type_key = @target_type_key.split '.'
|
175
|
-
end
|
176
|
-
|
177
|
-
if @remove_keys_on_update && @remove_keys_on_update.is_a?(String)
|
178
|
-
@remove_keys_on_update = @remove_keys_on_update.split ','
|
179
|
-
end
|
180
|
-
|
181
|
-
raise Fluent::ConfigError, "'max_retry_putting_template' must be positive number." if @max_retry_putting_template < 0
|
182
|
-
|
183
|
-
# Raise error when using host placeholders and template features at same time.
|
184
|
-
valid_host_placeholder = placeholder?(:host_placeholder, @host)
|
185
|
-
if valid_host_placeholder && (@template_name && @template_file || @templates)
|
186
|
-
raise Fluent::ConfigError, "host placeholder and template installation are exclusive features."
|
187
|
-
end
|
188
|
-
|
189
|
-
if !Fluent::Engine.dry_run_mode
|
190
|
-
if @template_name && @template_file
|
191
|
-
retry_operate(@max_retry_putting_template, @fail_on_putting_template_retry_exceed) do
|
192
|
-
if @customize_template
|
193
|
-
if @rollover_index
|
194
|
-
raise Fluent::ConfigError, "'deflector_alias' must be provided if 'rollover_index' is set true ." if not @deflector_alias
|
195
|
-
end
|
196
|
-
template_custom_install(@template_name, @template_file, @template_overwrite, @customize_template, @index_prefix, @rollover_index, @deflector_alias, @application_name, @index_date_pattern)
|
197
|
-
else
|
198
|
-
template_install(@template_name, @template_file, @template_overwrite)
|
199
|
-
end
|
200
|
-
end
|
201
|
-
elsif @templates
|
202
|
-
retry_operate(@max_retry_putting_template, @fail_on_putting_template_retry_exceed) do
|
203
|
-
templates_hash_install(@templates, @template_overwrite)
|
204
|
-
end
|
205
|
-
end
|
206
|
-
end
|
207
|
-
|
208
|
-
@serializer_class = nil
|
209
|
-
begin
|
210
|
-
require 'oj'
|
211
|
-
@dump_proc = Oj.method(:dump)
|
212
|
-
if @prefer_oj_serializer
|
213
|
-
@serializer_class = Fluent::Plugin::Serializer::Oj
|
214
|
-
Elasticsearch::API.settings[:serializer] = Fluent::Plugin::Serializer::Oj
|
215
|
-
end
|
216
|
-
rescue LoadError
|
217
|
-
@dump_proc = Yajl.method(:dump)
|
218
|
-
end
|
219
|
-
|
220
|
-
if @user && m = @user.match(/%{(?<user>.*)}/)
|
221
|
-
@user = URI.encode_www_form_component(m["user"])
|
222
|
-
end
|
223
|
-
if @password && m = @password.match(/%{(?<password>.*)}/)
|
224
|
-
@password = URI.encode_www_form_component(m["password"])
|
225
|
-
end
|
226
|
-
|
227
|
-
@transport_logger = nil
|
228
|
-
if @with_transporter_log
|
229
|
-
@transport_logger = log
|
230
|
-
log_level = conf['@log_level'] || conf['log_level']
|
231
|
-
log.warn "Consider to specify log_level with @log_level." unless log_level
|
232
|
-
end
|
233
|
-
# Specify @sniffer_class before calling #client.
|
234
|
-
# #detect_es_major_version uses #client.
|
235
|
-
@sniffer_class = nil
|
236
|
-
begin
|
237
|
-
@sniffer_class = Object.const_get(@sniffer_class_name) if @sniffer_class_name
|
238
|
-
rescue Exception => ex
|
239
|
-
raise Fluent::ConfigError, "Could not load sniffer class #{@sniffer_class_name}: #{ex}"
|
240
|
-
end
|
241
|
-
|
242
|
-
@last_seen_major_version =
|
243
|
-
if @verify_es_version_at_startup && !Fluent::Engine.dry_run_mode
|
244
|
-
retry_operate(@max_retry_get_es_version) do
|
245
|
-
detect_es_major_version
|
246
|
-
end
|
247
|
-
else
|
248
|
-
@default_elasticsearch_version
|
249
|
-
end
|
250
|
-
if @last_seen_major_version == 6 && @type_name != DEFAULT_TYPE_NAME_ES_7x
|
251
|
-
log.info "Detected ES 6.x: ES 7.x will only accept `_doc` in type_name."
|
252
|
-
end
|
253
|
-
if @last_seen_major_version >= 7 && @type_name != DEFAULT_TYPE_NAME_ES_7x
|
254
|
-
log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`."
|
255
|
-
@type_name = '_doc'.freeze
|
256
|
-
end
|
257
|
-
|
258
|
-
if @validate_client_version && !Fluent::Engine.dry_run_mode
|
259
|
-
if @last_seen_major_version != client_library_version.to_i
|
260
|
-
raise Fluent::ConfigError, <<-EOC
|
261
|
-
Detected ES #{@last_seen_major_version} but you use ES client #{client_library_version}.
|
262
|
-
Please consider to use #{@last_seen_major_version}.x series ES client.
|
263
|
-
EOC
|
264
|
-
end
|
265
|
-
end
|
266
|
-
|
267
|
-
if @last_seen_major_version >= 6
|
268
|
-
case @ssl_version
|
269
|
-
when :SSLv23, :TLSv1, :TLSv1_1
|
270
|
-
if @scheme == :https
|
271
|
-
log.warn "Detected ES 6.x or above and enabled insecure security:
|
272
|
-
You might have to specify `ssl_version TLSv1_2` in configuration."
|
273
|
-
end
|
274
|
-
end
|
275
|
-
end
|
276
|
-
|
277
|
-
if @buffer_config.flush_thread_count < 2
|
278
|
-
log.warn "To prevent events traffic jam, you should specify 2 or more 'flush_thread_count'."
|
279
|
-
end
|
280
|
-
|
281
|
-
# Consider missing the prefix of "$." in nested key specifiers.
|
282
|
-
@id_key = convert_compat_id_key(@id_key) if @id_key
|
283
|
-
@parent_key = convert_compat_id_key(@parent_key) if @parent_key
|
284
|
-
@routing_key = convert_compat_id_key(@routing_key) if @routing_key
|
285
|
-
|
286
|
-
@routing_key_name = configure_routing_key_name
|
287
|
-
@meta_config_map = create_meta_config_map
|
288
|
-
@current_config = nil
|
289
|
-
|
290
|
-
@ignore_exception_classes = @ignore_exceptions.map do |exception|
|
291
|
-
unless Object.const_defined?(exception)
|
292
|
-
log.warn "Cannot find class #{exception}. Will ignore it."
|
293
|
-
|
294
|
-
nil
|
295
|
-
else
|
296
|
-
Object.const_get(exception)
|
297
|
-
end
|
298
|
-
end.compact
|
299
|
-
|
300
|
-
if @bulk_message_request_threshold < 0
|
301
|
-
class << self
|
302
|
-
alias_method :split_request?, :split_request_size_uncheck?
|
303
|
-
end
|
304
|
-
else
|
305
|
-
class << self
|
306
|
-
alias_method :split_request?, :split_request_size_check?
|
307
|
-
end
|
308
|
-
end
|
309
|
-
end
|
310
|
-
|
311
|
-
def placeholder?(name, param)
|
312
|
-
begin
|
313
|
-
placeholder_validate!(name, param)
|
314
|
-
true
|
315
|
-
rescue Fluent::ConfigError
|
316
|
-
false
|
317
|
-
end
|
318
|
-
end
|
319
|
-
|
320
|
-
def backend_options
|
321
|
-
case @http_backend
|
322
|
-
when :excon
|
323
|
-
{ client_key: @client_key, client_cert: @client_cert, client_key_pass: @client_key_pass }
|
324
|
-
when :typhoeus
|
325
|
-
require 'typhoeus'
|
326
|
-
{ sslkey: @client_key, sslcert: @client_cert, keypasswd: @client_key_pass }
|
327
|
-
end
|
328
|
-
rescue LoadError => ex
|
329
|
-
log.error_backtrace(ex.backtrace)
|
330
|
-
raise Fluent::ConfigError, "You must install #{@http_backend} gem. Exception: #{ex}"
|
331
|
-
end
|
332
|
-
|
333
|
-
def detect_es_major_version
|
334
|
-
@_es_info ||= client.info
|
335
|
-
@_es_info["version"]["number"].to_i
|
336
|
-
end
|
337
|
-
|
338
|
-
def client_library_version
|
339
|
-
Elasticsearch::VERSION
|
340
|
-
end
|
341
|
-
|
342
|
-
def configure_routing_key_name
|
343
|
-
if @last_seen_major_version >= 7
|
344
|
-
'routing'
|
345
|
-
else
|
346
|
-
'_routing'
|
347
|
-
end
|
348
|
-
end
|
349
|
-
|
350
|
-
def convert_compat_id_key(key)
|
351
|
-
if key.include?('.') && !key.start_with?('$[')
|
352
|
-
key = "$.#{key}" unless key.start_with?('$.')
|
353
|
-
end
|
354
|
-
key
|
355
|
-
end
|
356
|
-
|
357
|
-
def create_meta_config_map
|
358
|
-
result = []
|
359
|
-
result << [record_accessor_create(@id_key), '_id'] if @id_key
|
360
|
-
result << [record_accessor_create(@parent_key), '_parent'] if @parent_key
|
361
|
-
result << [record_accessor_create(@routing_key), @routing_key_name] if @routing_key
|
362
|
-
result
|
363
|
-
end
|
364
|
-
|
365
|
-
# once fluent v0.14 is released we might be able to use
|
366
|
-
# Fluent::Parser::TimeParser, but it doesn't quite do what we want - if gives
|
367
|
-
# [sec,nsec] where as we want something we can call `strftime` on...
|
368
|
-
def create_time_parser
|
369
|
-
if @time_key_format
|
370
|
-
begin
|
371
|
-
# Strptime doesn't support all formats, but for those it does it's
|
372
|
-
# blazingly fast.
|
373
|
-
strptime = Strptime.new(@time_key_format)
|
374
|
-
Proc.new { |value|
|
375
|
-
value = convert_numeric_time_into_string(value, @time_key_format) if value.is_a?(Numeric)
|
376
|
-
strptime.exec(value).to_datetime
|
377
|
-
}
|
378
|
-
rescue
|
379
|
-
# Can happen if Strptime doesn't recognize the format; or
|
380
|
-
# if strptime couldn't be required (because it's not installed -- it's
|
381
|
-
# ruby 2 only)
|
382
|
-
Proc.new { |value|
|
383
|
-
value = convert_numeric_time_into_string(value, @time_key_format) if value.is_a?(Numeric)
|
384
|
-
DateTime.strptime(value, @time_key_format)
|
385
|
-
}
|
386
|
-
end
|
387
|
-
else
|
388
|
-
Proc.new { |value|
|
389
|
-
value = convert_numeric_time_into_string(value) if value.is_a?(Numeric)
|
390
|
-
DateTime.parse(value)
|
391
|
-
}
|
392
|
-
end
|
393
|
-
end
|
394
|
-
|
395
|
-
def convert_numeric_time_into_string(numeric_time, time_key_format = "%Y-%m-%d %H:%M:%S.%N%z")
|
396
|
-
numeric_time_parser = Fluent::NumericTimeParser.new(:float)
|
397
|
-
Time.at(numeric_time_parser.parse(numeric_time).to_r).strftime(time_key_format)
|
398
|
-
end
|
399
|
-
|
400
|
-
def parse_time(value, event_time, tag)
|
401
|
-
@time_parser.call(value)
|
402
|
-
rescue => e
|
403
|
-
router.emit_error_event(@time_parse_error_tag, Fluent::Engine.now, {'tag' => tag, 'time' => event_time, 'format' => @time_key_format, 'value' => value}, e)
|
404
|
-
return Time.at(event_time).to_datetime
|
405
|
-
end
|
406
|
-
|
407
|
-
def client(host = nil)
|
408
|
-
# check here to see if we already have a client connection for the given host
|
409
|
-
connection_options = get_connection_options(host)
|
410
|
-
|
411
|
-
@_es = nil unless is_existing_connection(connection_options[:hosts])
|
412
|
-
|
413
|
-
@_es ||= begin
|
414
|
-
@current_config = connection_options[:hosts].clone
|
415
|
-
adapter_conf = lambda {|f| f.adapter @http_backend, @backend_options }
|
416
|
-
local_reload_connections = @reload_connections
|
417
|
-
if local_reload_connections && @reload_after > DEFAULT_RELOAD_AFTER
|
418
|
-
local_reload_connections = @reload_after
|
419
|
-
end
|
420
|
-
headers = { 'Content-Type' => @content_type.to_s }.merge(@custom_headers)
|
421
|
-
transport = Elasticsearch::Transport::Transport::HTTP::Faraday.new(connection_options.merge(
|
422
|
-
options: {
|
423
|
-
reload_connections: local_reload_connections,
|
424
|
-
reload_on_failure: @reload_on_failure,
|
425
|
-
resurrect_after: @resurrect_after,
|
426
|
-
logger: @transport_logger,
|
427
|
-
transport_options: {
|
428
|
-
headers: headers,
|
429
|
-
request: { timeout: @request_timeout },
|
430
|
-
ssl: { verify: @ssl_verify, ca_file: @ca_file, version: @ssl_version }
|
431
|
-
},
|
432
|
-
http: {
|
433
|
-
user: @user,
|
434
|
-
password: @password
|
435
|
-
},
|
436
|
-
sniffer_class: @sniffer_class,
|
437
|
-
serializer_class: @serializer_class,
|
438
|
-
}), &adapter_conf)
|
439
|
-
Elasticsearch::Client.new transport: transport
|
440
|
-
end
|
441
|
-
end
|
442
|
-
|
443
|
-
def get_escaped_userinfo(host_str)
|
444
|
-
if m = host_str.match(/(?<scheme>.*)%{(?<user>.*)}:%{(?<password>.*)}(?<path>@.*)/)
|
445
|
-
m["scheme"] +
|
446
|
-
URI.encode_www_form_component(m["user"]) +
|
447
|
-
':' +
|
448
|
-
URI.encode_www_form_component(m["password"]) +
|
449
|
-
m["path"]
|
450
|
-
else
|
451
|
-
host_str
|
452
|
-
end
|
453
|
-
end
|
454
|
-
|
455
|
-
def get_connection_options(con_host=nil)
|
456
|
-
raise "`password` must be present if `user` is present" if @user && !@password
|
457
|
-
|
458
|
-
hosts = if con_host || @hosts
|
459
|
-
(con_host || @hosts).split(',').map do |host_str|
|
460
|
-
# Support legacy hosts format host:port,host:port,host:port...
|
461
|
-
if host_str.match(%r{^[^:]+(\:\d+)?$})
|
462
|
-
{
|
463
|
-
host: host_str.split(':')[0],
|
464
|
-
port: (host_str.split(':')[1] || @port).to_i,
|
465
|
-
scheme: @scheme.to_s
|
466
|
-
}
|
467
|
-
else
|
468
|
-
# New hosts format expects URLs such as http://logs.foo.com,https://john:pass@logs2.foo.com/elastic
|
469
|
-
uri = URI(get_escaped_userinfo(host_str))
|
470
|
-
%w(user password path).inject(host: uri.host, port: uri.port, scheme: uri.scheme) do |hash, key|
|
471
|
-
hash[key.to_sym] = uri.public_send(key) unless uri.public_send(key).nil? || uri.public_send(key) == ''
|
472
|
-
hash
|
473
|
-
end
|
474
|
-
end
|
475
|
-
end.compact
|
476
|
-
else
|
477
|
-
[{host: @host, port: @port, scheme: @scheme.to_s}]
|
478
|
-
end.each do |host|
|
479
|
-
host.merge!(user: @user, password: @password) if !host[:user] && @user
|
480
|
-
host.merge!(path: @path) if !host[:path] && @path
|
481
|
-
end
|
482
|
-
|
483
|
-
{
|
484
|
-
hosts: hosts
|
485
|
-
}
|
486
|
-
end
|
487
|
-
|
488
|
-
def connection_options_description(con_host=nil)
|
489
|
-
get_connection_options(con_host)[:hosts].map do |host_info|
|
490
|
-
attributes = host_info.dup
|
491
|
-
attributes[:password] = 'obfuscated' if attributes.has_key?(:password)
|
492
|
-
attributes.inspect
|
493
|
-
end.join(', ')
|
494
|
-
end
|
495
|
-
|
496
|
-
# append_record_to_messages adds a record to the bulk message
|
497
|
-
# payload to be submitted to Elasticsearch. Records that do
|
498
|
-
# not include '_id' field are skipped when 'write_operation'
|
499
|
-
# is configured for 'create' or 'update'
|
500
|
-
#
|
501
|
-
# returns 'true' if record was appended to the bulk message
|
502
|
-
# and 'false' otherwise
|
503
|
-
def append_record_to_messages(op, meta, header, record, msgs)
|
504
|
-
case op
|
505
|
-
when UPDATE_OP, UPSERT_OP
|
506
|
-
if meta.has_key?(ID_FIELD)
|
507
|
-
header[UPDATE_OP] = meta
|
508
|
-
msgs << @dump_proc.call(header) << BODY_DELIMITER
|
509
|
-
msgs << @dump_proc.call(update_body(record, op)) << BODY_DELIMITER
|
510
|
-
return true
|
511
|
-
end
|
512
|
-
when CREATE_OP
|
513
|
-
if meta.has_key?(ID_FIELD)
|
514
|
-
header[CREATE_OP] = meta
|
515
|
-
msgs << @dump_proc.call(header) << BODY_DELIMITER
|
516
|
-
msgs << @dump_proc.call(record) << BODY_DELIMITER
|
517
|
-
return true
|
518
|
-
end
|
519
|
-
when INDEX_OP
|
520
|
-
header[INDEX_OP] = meta
|
521
|
-
msgs << @dump_proc.call(header) << BODY_DELIMITER
|
522
|
-
msgs << @dump_proc.call(record) << BODY_DELIMITER
|
523
|
-
return true
|
524
|
-
end
|
525
|
-
return false
|
526
|
-
end
|
527
|
-
|
528
|
-
def update_body(record, op)
|
529
|
-
update = remove_keys(record)
|
530
|
-
if @suppress_doc_wrap
|
531
|
-
return update
|
532
|
-
end
|
533
|
-
body = {"doc".freeze => update}
|
534
|
-
if op == UPSERT_OP
|
535
|
-
if update == record
|
536
|
-
body["doc_as_upsert".freeze] = true
|
537
|
-
else
|
538
|
-
body[UPSERT_OP] = record
|
539
|
-
end
|
540
|
-
end
|
541
|
-
body
|
542
|
-
end
|
543
|
-
|
544
|
-
def remove_keys(record)
|
545
|
-
keys = record[@remove_keys_on_update_key] || @remove_keys_on_update || []
|
546
|
-
record.delete(@remove_keys_on_update_key)
|
547
|
-
return record unless keys.any?
|
548
|
-
record = record.dup
|
549
|
-
keys.each { |key| record.delete(key) }
|
550
|
-
record
|
551
|
-
end
|
552
|
-
|
553
|
-
def flatten_record(record, prefix=[])
|
554
|
-
ret = {}
|
555
|
-
if record.is_a? Hash
|
556
|
-
record.each { |key, value|
|
557
|
-
ret.merge! flatten_record(value, prefix + [key.to_s])
|
558
|
-
}
|
559
|
-
elsif record.is_a? Array
|
560
|
-
# Don't mess with arrays, leave them unprocessed
|
561
|
-
ret.merge!({prefix.join(@flatten_hashes_separator) => record})
|
562
|
-
else
|
563
|
-
return {prefix.join(@flatten_hashes_separator) => record}
|
564
|
-
end
|
565
|
-
ret
|
566
|
-
end
|
567
|
-
|
568
|
-
def expand_placeholders(chunk)
|
569
|
-
logstash_prefix = extract_placeholders(@logstash_prefix, chunk)
|
570
|
-
index_name = extract_placeholders(@index_name, chunk)
|
571
|
-
type_name = extract_placeholders(@type_name, chunk)
|
572
|
-
return logstash_prefix, index_name, type_name
|
573
|
-
end
|
574
|
-
|
575
|
-
def multi_workers_ready?
|
576
|
-
true
|
577
|
-
end
|
578
|
-
|
579
|
-
def write(chunk)
|
580
|
-
bulk_message_count = Hash.new { |h,k| h[k] = 0 }
|
581
|
-
bulk_message = Hash.new { |h,k| h[k] = '' }
|
582
|
-
header = {}
|
583
|
-
meta = {}
|
584
|
-
|
585
|
-
tag = chunk.metadata.tag
|
586
|
-
extracted_values = expand_placeholders(chunk)
|
587
|
-
host = if @hosts
|
588
|
-
extract_placeholders(@hosts, chunk)
|
589
|
-
else
|
590
|
-
extract_placeholders(@host, chunk)
|
591
|
-
end
|
592
|
-
|
593
|
-
chunk.msgpack_each do |time, record|
|
594
|
-
next unless record.is_a? Hash
|
595
|
-
begin
|
596
|
-
meta, header, record = process_message(tag, meta, header, time, record, extracted_values)
|
597
|
-
info = if @include_index_in_url
|
598
|
-
RequestInfo.new(host, meta.delete("_index".freeze))
|
599
|
-
else
|
600
|
-
RequestInfo.new(host, nil)
|
601
|
-
end
|
602
|
-
|
603
|
-
if
|
604
|
-
|
605
|
-
|
606
|
-
|
607
|
-
|
608
|
-
|
609
|
-
|
610
|
-
|
611
|
-
|
612
|
-
|
613
|
-
|
614
|
-
|
615
|
-
|
616
|
-
|
617
|
-
|
618
|
-
|
619
|
-
|
620
|
-
|
621
|
-
|
622
|
-
|
623
|
-
|
624
|
-
|
625
|
-
|
626
|
-
|
627
|
-
|
628
|
-
msgs.
|
629
|
-
|
630
|
-
|
631
|
-
|
632
|
-
|
633
|
-
|
634
|
-
|
635
|
-
|
636
|
-
|
637
|
-
|
638
|
-
|
639
|
-
|
640
|
-
|
641
|
-
|
642
|
-
|
643
|
-
|
644
|
-
|
645
|
-
|
646
|
-
|
647
|
-
|
648
|
-
|
649
|
-
|
650
|
-
|
651
|
-
|
652
|
-
|
653
|
-
|
654
|
-
|
655
|
-
|
656
|
-
|
657
|
-
|
658
|
-
|
659
|
-
|
660
|
-
|
661
|
-
|
662
|
-
|
663
|
-
|
664
|
-
|
665
|
-
|
666
|
-
|
667
|
-
|
668
|
-
|
669
|
-
|
670
|
-
|
671
|
-
|
672
|
-
|
673
|
-
|
674
|
-
|
675
|
-
|
676
|
-
|
677
|
-
#
|
678
|
-
|
679
|
-
|
680
|
-
|
681
|
-
|
682
|
-
|
683
|
-
|
684
|
-
|
685
|
-
|
686
|
-
|
687
|
-
|
688
|
-
|
689
|
-
|
690
|
-
|
691
|
-
|
692
|
-
|
693
|
-
|
694
|
-
|
695
|
-
|
696
|
-
|
697
|
-
|
698
|
-
|
699
|
-
|
700
|
-
|
701
|
-
|
702
|
-
|
703
|
-
meta
|
704
|
-
meta["
|
705
|
-
|
706
|
-
|
707
|
-
|
708
|
-
|
709
|
-
|
710
|
-
|
711
|
-
|
712
|
-
|
713
|
-
|
714
|
-
|
715
|
-
|
716
|
-
|
717
|
-
|
718
|
-
|
719
|
-
|
720
|
-
|
721
|
-
|
722
|
-
|
723
|
-
|
724
|
-
# returns [
|
725
|
-
|
726
|
-
|
727
|
-
|
728
|
-
|
729
|
-
|
730
|
-
|
731
|
-
#
|
732
|
-
|
733
|
-
|
734
|
-
|
735
|
-
|
736
|
-
|
737
|
-
|
738
|
-
|
739
|
-
|
740
|
-
|
741
|
-
error.
|
742
|
-
|
743
|
-
|
744
|
-
|
745
|
-
|
746
|
-
|
747
|
-
|
748
|
-
|
749
|
-
|
750
|
-
|
751
|
-
|
752
|
-
@
|
753
|
-
|
754
|
-
|
755
|
-
|
756
|
-
|
757
|
-
|
758
|
-
|
759
|
-
|
760
|
-
|
761
|
-
|
762
|
-
|
763
|
-
|
764
|
-
return false if @
|
765
|
-
return false if
|
766
|
-
|
767
|
-
|
768
|
-
|
769
|
-
|
770
|
-
|
771
|
-
|
772
|
-
|
773
|
-
|
774
|
-
|
775
|
-
|
776
|
-
end
|
1
|
+
# encoding: UTF-8
|
2
|
+
require 'date'
|
3
|
+
require 'excon'
|
4
|
+
require 'elasticsearch'
|
5
|
+
require 'json'
|
6
|
+
require 'uri'
|
7
|
+
begin
|
8
|
+
require 'strptime'
|
9
|
+
rescue LoadError
|
10
|
+
end
|
11
|
+
|
12
|
+
require 'fluent/plugin/output'
|
13
|
+
require 'fluent/event'
|
14
|
+
require 'fluent/error'
|
15
|
+
require 'fluent/time'
|
16
|
+
require 'fluent/log-ext'
|
17
|
+
require_relative 'elasticsearch_constants'
|
18
|
+
require_relative 'elasticsearch_error'
|
19
|
+
require_relative 'elasticsearch_error_handler'
|
20
|
+
require_relative 'elasticsearch_index_template'
|
21
|
+
begin
|
22
|
+
require_relative 'oj_serializer'
|
23
|
+
rescue LoadError
|
24
|
+
end
|
25
|
+
|
26
|
+
module Fluent::Plugin
|
27
|
+
class ElasticsearchOutput < Output
|
28
|
+
class RecoverableRequestFailure < StandardError; end
|
29
|
+
class UnrecoverableRequestFailure < Fluent::UnrecoverableError; end
|
30
|
+
|
31
|
+
# MissingIdFieldError is raised for records that do not
|
32
|
+
# include the field for the unique record identifier
|
33
|
+
class MissingIdFieldError < StandardError; end
|
34
|
+
|
35
|
+
# RetryStreamError privides a stream to be
|
36
|
+
# put back in the pipeline for cases where a bulk request
|
37
|
+
# failed (e.g some records succeed while others failed)
|
38
|
+
class RetryStreamError < StandardError
|
39
|
+
attr_reader :retry_stream
|
40
|
+
def initialize(retry_stream)
|
41
|
+
@retry_stream = retry_stream
|
42
|
+
end
|
43
|
+
end
|
44
|
+
|
45
|
+
RequestInfo = Struct.new(:host, :index)
|
46
|
+
|
47
|
+
helpers :event_emitter, :compat_parameters, :record_accessor
|
48
|
+
|
49
|
+
Fluent::Plugin.register_output('elasticsearch', self)
|
50
|
+
|
51
|
+
DEFAULT_BUFFER_TYPE = "memory"
|
52
|
+
DEFAULT_ELASTICSEARCH_VERSION = 5 # For compatibility.
|
53
|
+
DEFAULT_TYPE_NAME_ES_7x = "_doc".freeze
|
54
|
+
DEFAULT_TYPE_NAME = "fluentd".freeze
|
55
|
+
DEFAULT_RELOAD_AFTER = -1
|
56
|
+
TARGET_BULK_BYTES = 20 * 1024 * 1024
|
57
|
+
|
58
|
+
config_param :host, :string, :default => 'localhost'
|
59
|
+
config_param :port, :integer, :default => 9200
|
60
|
+
config_param :user, :string, :default => nil
|
61
|
+
config_param :password, :string, :default => nil, :secret => true
|
62
|
+
config_param :path, :string, :default => nil
|
63
|
+
config_param :scheme, :enum, :list => [:https, :http], :default => :http
|
64
|
+
config_param :hosts, :string, :default => nil
|
65
|
+
config_param :target_index_key, :string, :default => nil
|
66
|
+
config_param :target_type_key, :string, :default => nil,
|
67
|
+
:deprecated => <<EOC
|
68
|
+
Elasticsearch 7.x or above will ignore this config. Please use fixed type_name instead.
|
69
|
+
EOC
|
70
|
+
config_param :time_key_format, :string, :default => nil
|
71
|
+
config_param :time_precision, :integer, :default => 9
|
72
|
+
config_param :include_timestamp, :bool, :default => false
|
73
|
+
config_param :logstash_format, :bool, :default => false
|
74
|
+
config_param :logstash_prefix, :string, :default => "logstash"
|
75
|
+
config_param :logstash_prefix_separator, :string, :default => '-'
|
76
|
+
config_param :logstash_dateformat, :string, :default => "%Y.%m.%d"
|
77
|
+
config_param :utc_index, :bool, :default => true
|
78
|
+
config_param :type_name, :string, :default => DEFAULT_TYPE_NAME
|
79
|
+
config_param :index_name, :string, :default => "fluentd"
|
80
|
+
config_param :id_key, :string, :default => nil
|
81
|
+
config_param :write_operation, :string, :default => "index"
|
82
|
+
config_param :parent_key, :string, :default => nil
|
83
|
+
config_param :routing_key, :string, :default => nil
|
84
|
+
config_param :request_timeout, :time, :default => 5
|
85
|
+
config_param :reload_connections, :bool, :default => true
|
86
|
+
config_param :reload_on_failure, :bool, :default => false
|
87
|
+
config_param :retry_tag, :string, :default=>nil
|
88
|
+
config_param :resurrect_after, :time, :default => 60
|
89
|
+
config_param :time_key, :string, :default => nil
|
90
|
+
config_param :time_key_exclude_timestamp, :bool, :default => false
|
91
|
+
config_param :ssl_verify , :bool, :default => true
|
92
|
+
config_param :client_key, :string, :default => nil
|
93
|
+
config_param :client_cert, :string, :default => nil
|
94
|
+
config_param :client_key_pass, :string, :default => nil
|
95
|
+
config_param :ca_file, :string, :default => nil
|
96
|
+
config_param :ssl_version, :enum, list: [:SSLv23, :TLSv1, :TLSv1_1, :TLSv1_2], :default => :TLSv1
|
97
|
+
config_param :remove_keys, :string, :default => nil
|
98
|
+
config_param :remove_keys_on_update, :string, :default => ""
|
99
|
+
config_param :remove_keys_on_update_key, :string, :default => nil
|
100
|
+
config_param :flatten_hashes, :bool, :default => false
|
101
|
+
config_param :flatten_hashes_separator, :string, :default => "_"
|
102
|
+
config_param :template_name, :string, :default => nil
|
103
|
+
config_param :template_file, :string, :default => nil
|
104
|
+
config_param :template_overwrite, :bool, :default => false
|
105
|
+
config_param :customize_template, :hash, :default => nil
|
106
|
+
config_param :rollover_index, :string, :default => false
|
107
|
+
config_param :index_date_pattern, :string, :default => "now/d"
|
108
|
+
config_param :deflector_alias, :string, :default => nil
|
109
|
+
config_param :index_prefix, :string, :default => "logstash"
|
110
|
+
config_param :application_name, :string, :default => "default"
|
111
|
+
config_param :templates, :hash, :default => nil
|
112
|
+
config_param :max_retry_putting_template, :integer, :default => 10
|
113
|
+
config_param :fail_on_putting_template_retry_exceed, :bool, :default => true
|
114
|
+
config_param :max_retry_get_es_version, :integer, :default => 15
|
115
|
+
config_param :include_tag_key, :bool, :default => false
|
116
|
+
config_param :tag_key, :string, :default => 'tag'
|
117
|
+
config_param :time_parse_error_tag, :string, :default => 'Fluent::ElasticsearchOutput::TimeParser.error'
|
118
|
+
config_param :reconnect_on_error, :bool, :default => false
|
119
|
+
config_param :pipeline, :string, :default => nil
|
120
|
+
config_param :with_transporter_log, :bool, :default => false
|
121
|
+
config_param :emit_error_for_missing_id, :bool, :default => false
|
122
|
+
config_param :sniffer_class_name, :string, :default => nil
|
123
|
+
config_param :reload_after, :integer, :default => DEFAULT_RELOAD_AFTER
|
124
|
+
config_param :content_type, :enum, list: [:"application/json", :"application/x-ndjson"], :default => :"application/json",
|
125
|
+
:deprecated => <<EOC
|
126
|
+
elasticsearch gem v6.0.2 starts to use correct Content-Type. Please upgrade elasticserach gem and stop to use this option.
|
127
|
+
see: https://github.com/elastic/elasticsearch-ruby/pull/514
|
128
|
+
EOC
|
129
|
+
config_param :include_index_in_url, :bool, :default => false
|
130
|
+
config_param :http_backend, :enum, list: [:excon, :typhoeus], :default => :excon
|
131
|
+
config_param :validate_client_version, :bool, :default => false
|
132
|
+
config_param :prefer_oj_serializer, :bool, :default => false
|
133
|
+
config_param :unrecoverable_error_types, :array, :default => ["out_of_memory_error", "es_rejected_execution_exception"]
|
134
|
+
config_param :verify_es_version_at_startup, :bool, :default => true
|
135
|
+
config_param :default_elasticsearch_version, :integer, :default => DEFAULT_ELASTICSEARCH_VERSION
|
136
|
+
config_param :log_es_400_reason, :bool, :default => false
|
137
|
+
config_param :custom_headers, :hash, :default => {}
|
138
|
+
config_param :suppress_doc_wrap, :bool, :default => false
|
139
|
+
config_param :ignore_exceptions, :array, :default => [], value_type: :string, :desc => "Ignorable exception list"
|
140
|
+
config_param :exception_backup, :bool, :default => true, :desc => "Chunk backup flag when ignore exception occured"
|
141
|
+
config_param :bulk_message_request_threshold, :size, :default => TARGET_BULK_BYTES
|
142
|
+
|
143
|
+
config_section :buffer do
|
144
|
+
config_set_default :@type, DEFAULT_BUFFER_TYPE
|
145
|
+
config_set_default :chunk_keys, ['tag']
|
146
|
+
config_set_default :timekey_use_utc, true
|
147
|
+
end
|
148
|
+
|
149
|
+
include Fluent::ElasticsearchIndexTemplate
|
150
|
+
include Fluent::Plugin::ElasticsearchConstants
|
151
|
+
|
152
|
+
def initialize
|
153
|
+
super
|
154
|
+
end
|
155
|
+
|
156
|
+
def configure(conf)
|
157
|
+
compat_parameters_convert(conf, :buffer)
|
158
|
+
|
159
|
+
super
|
160
|
+
raise Fluent::ConfigError, "'tag' in chunk_keys is required." if not @chunk_key_tag
|
161
|
+
|
162
|
+
@time_parser = create_time_parser
|
163
|
+
@backend_options = backend_options
|
164
|
+
|
165
|
+
if @remove_keys
|
166
|
+
@remove_keys = @remove_keys.split(/\s*,\s*/)
|
167
|
+
end
|
168
|
+
|
169
|
+
if @target_index_key && @target_index_key.is_a?(String)
|
170
|
+
@target_index_key = @target_index_key.split '.'
|
171
|
+
end
|
172
|
+
|
173
|
+
if @target_type_key && @target_type_key.is_a?(String)
|
174
|
+
@target_type_key = @target_type_key.split '.'
|
175
|
+
end
|
176
|
+
|
177
|
+
if @remove_keys_on_update && @remove_keys_on_update.is_a?(String)
|
178
|
+
@remove_keys_on_update = @remove_keys_on_update.split ','
|
179
|
+
end
|
180
|
+
|
181
|
+
raise Fluent::ConfigError, "'max_retry_putting_template' must be positive number." if @max_retry_putting_template < 0
|
182
|
+
|
183
|
+
# Raise error when using host placeholders and template features at same time.
|
184
|
+
valid_host_placeholder = placeholder?(:host_placeholder, @host)
|
185
|
+
if valid_host_placeholder && (@template_name && @template_file || @templates)
|
186
|
+
raise Fluent::ConfigError, "host placeholder and template installation are exclusive features."
|
187
|
+
end
|
188
|
+
|
189
|
+
if !Fluent::Engine.dry_run_mode
|
190
|
+
if @template_name && @template_file
|
191
|
+
retry_operate(@max_retry_putting_template, @fail_on_putting_template_retry_exceed) do
|
192
|
+
if @customize_template
|
193
|
+
if @rollover_index
|
194
|
+
raise Fluent::ConfigError, "'deflector_alias' must be provided if 'rollover_index' is set true ." if not @deflector_alias
|
195
|
+
end
|
196
|
+
template_custom_install(@template_name, @template_file, @template_overwrite, @customize_template, @index_prefix, @rollover_index, @deflector_alias, @application_name, @index_date_pattern)
|
197
|
+
else
|
198
|
+
template_install(@template_name, @template_file, @template_overwrite)
|
199
|
+
end
|
200
|
+
end
|
201
|
+
elsif @templates
|
202
|
+
retry_operate(@max_retry_putting_template, @fail_on_putting_template_retry_exceed) do
|
203
|
+
templates_hash_install(@templates, @template_overwrite)
|
204
|
+
end
|
205
|
+
end
|
206
|
+
end
|
207
|
+
|
208
|
+
@serializer_class = nil
|
209
|
+
begin
|
210
|
+
require 'oj'
|
211
|
+
@dump_proc = Oj.method(:dump)
|
212
|
+
if @prefer_oj_serializer
|
213
|
+
@serializer_class = Fluent::Plugin::Serializer::Oj
|
214
|
+
Elasticsearch::API.settings[:serializer] = Fluent::Plugin::Serializer::Oj
|
215
|
+
end
|
216
|
+
rescue LoadError
|
217
|
+
@dump_proc = Yajl.method(:dump)
|
218
|
+
end
|
219
|
+
|
220
|
+
if @user && m = @user.match(/%{(?<user>.*)}/)
|
221
|
+
@user = URI.encode_www_form_component(m["user"])
|
222
|
+
end
|
223
|
+
if @password && m = @password.match(/%{(?<password>.*)}/)
|
224
|
+
@password = URI.encode_www_form_component(m["password"])
|
225
|
+
end
|
226
|
+
|
227
|
+
@transport_logger = nil
|
228
|
+
if @with_transporter_log
|
229
|
+
@transport_logger = log
|
230
|
+
log_level = conf['@log_level'] || conf['log_level']
|
231
|
+
log.warn "Consider to specify log_level with @log_level." unless log_level
|
232
|
+
end
|
233
|
+
# Specify @sniffer_class before calling #client.
|
234
|
+
# #detect_es_major_version uses #client.
|
235
|
+
@sniffer_class = nil
|
236
|
+
begin
|
237
|
+
@sniffer_class = Object.const_get(@sniffer_class_name) if @sniffer_class_name
|
238
|
+
rescue Exception => ex
|
239
|
+
raise Fluent::ConfigError, "Could not load sniffer class #{@sniffer_class_name}: #{ex}"
|
240
|
+
end
|
241
|
+
|
242
|
+
@last_seen_major_version =
|
243
|
+
if @verify_es_version_at_startup && !Fluent::Engine.dry_run_mode
|
244
|
+
retry_operate(@max_retry_get_es_version) do
|
245
|
+
detect_es_major_version
|
246
|
+
end
|
247
|
+
else
|
248
|
+
@default_elasticsearch_version
|
249
|
+
end
|
250
|
+
if @last_seen_major_version == 6 && @type_name != DEFAULT_TYPE_NAME_ES_7x
|
251
|
+
log.info "Detected ES 6.x: ES 7.x will only accept `_doc` in type_name."
|
252
|
+
end
|
253
|
+
if @last_seen_major_version >= 7 && @type_name != DEFAULT_TYPE_NAME_ES_7x
|
254
|
+
log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`."
|
255
|
+
@type_name = '_doc'.freeze
|
256
|
+
end
|
257
|
+
|
258
|
+
if @validate_client_version && !Fluent::Engine.dry_run_mode
|
259
|
+
if @last_seen_major_version != client_library_version.to_i
|
260
|
+
raise Fluent::ConfigError, <<-EOC
|
261
|
+
Detected ES #{@last_seen_major_version} but you use ES client #{client_library_version}.
|
262
|
+
Please consider to use #{@last_seen_major_version}.x series ES client.
|
263
|
+
EOC
|
264
|
+
end
|
265
|
+
end
|
266
|
+
|
267
|
+
if @last_seen_major_version >= 6
|
268
|
+
case @ssl_version
|
269
|
+
when :SSLv23, :TLSv1, :TLSv1_1
|
270
|
+
if @scheme == :https
|
271
|
+
log.warn "Detected ES 6.x or above and enabled insecure security:
|
272
|
+
You might have to specify `ssl_version TLSv1_2` in configuration."
|
273
|
+
end
|
274
|
+
end
|
275
|
+
end
|
276
|
+
|
277
|
+
if @buffer_config.flush_thread_count < 2
|
278
|
+
log.warn "To prevent events traffic jam, you should specify 2 or more 'flush_thread_count'."
|
279
|
+
end
|
280
|
+
|
281
|
+
# Consider missing the prefix of "$." in nested key specifiers.
|
282
|
+
@id_key = convert_compat_id_key(@id_key) if @id_key
|
283
|
+
@parent_key = convert_compat_id_key(@parent_key) if @parent_key
|
284
|
+
@routing_key = convert_compat_id_key(@routing_key) if @routing_key
|
285
|
+
|
286
|
+
@routing_key_name = configure_routing_key_name
|
287
|
+
@meta_config_map = create_meta_config_map
|
288
|
+
@current_config = nil
|
289
|
+
|
290
|
+
@ignore_exception_classes = @ignore_exceptions.map do |exception|
|
291
|
+
unless Object.const_defined?(exception)
|
292
|
+
log.warn "Cannot find class #{exception}. Will ignore it."
|
293
|
+
|
294
|
+
nil
|
295
|
+
else
|
296
|
+
Object.const_get(exception)
|
297
|
+
end
|
298
|
+
end.compact
|
299
|
+
|
300
|
+
if @bulk_message_request_threshold < 0
|
301
|
+
class << self
|
302
|
+
alias_method :split_request?, :split_request_size_uncheck?
|
303
|
+
end
|
304
|
+
else
|
305
|
+
class << self
|
306
|
+
alias_method :split_request?, :split_request_size_check?
|
307
|
+
end
|
308
|
+
end
|
309
|
+
end
|
310
|
+
|
311
|
+
def placeholder?(name, param)
|
312
|
+
begin
|
313
|
+
placeholder_validate!(name, param)
|
314
|
+
true
|
315
|
+
rescue Fluent::ConfigError
|
316
|
+
false
|
317
|
+
end
|
318
|
+
end
|
319
|
+
|
320
|
+
def backend_options
|
321
|
+
case @http_backend
|
322
|
+
when :excon
|
323
|
+
{ client_key: @client_key, client_cert: @client_cert, client_key_pass: @client_key_pass }
|
324
|
+
when :typhoeus
|
325
|
+
require 'typhoeus'
|
326
|
+
{ sslkey: @client_key, sslcert: @client_cert, keypasswd: @client_key_pass }
|
327
|
+
end
|
328
|
+
rescue LoadError => ex
|
329
|
+
log.error_backtrace(ex.backtrace)
|
330
|
+
raise Fluent::ConfigError, "You must install #{@http_backend} gem. Exception: #{ex}"
|
331
|
+
end
|
332
|
+
|
333
|
+
def detect_es_major_version
|
334
|
+
@_es_info ||= client.info
|
335
|
+
@_es_info["version"]["number"].to_i
|
336
|
+
end
|
337
|
+
|
338
|
+
def client_library_version
|
339
|
+
Elasticsearch::VERSION
|
340
|
+
end
|
341
|
+
|
342
|
+
def configure_routing_key_name
|
343
|
+
if @last_seen_major_version >= 7
|
344
|
+
'routing'
|
345
|
+
else
|
346
|
+
'_routing'
|
347
|
+
end
|
348
|
+
end
|
349
|
+
|
350
|
+
def convert_compat_id_key(key)
|
351
|
+
if key.include?('.') && !key.start_with?('$[')
|
352
|
+
key = "$.#{key}" unless key.start_with?('$.')
|
353
|
+
end
|
354
|
+
key
|
355
|
+
end
|
356
|
+
|
357
|
+
def create_meta_config_map
|
358
|
+
result = []
|
359
|
+
result << [record_accessor_create(@id_key), '_id'] if @id_key
|
360
|
+
result << [record_accessor_create(@parent_key), '_parent'] if @parent_key
|
361
|
+
result << [record_accessor_create(@routing_key), @routing_key_name] if @routing_key
|
362
|
+
result
|
363
|
+
end
|
364
|
+
|
365
|
+
# once fluent v0.14 is released we might be able to use
|
366
|
+
# Fluent::Parser::TimeParser, but it doesn't quite do what we want - if gives
|
367
|
+
# [sec,nsec] where as we want something we can call `strftime` on...
|
368
|
+
def create_time_parser
|
369
|
+
if @time_key_format
|
370
|
+
begin
|
371
|
+
# Strptime doesn't support all formats, but for those it does it's
|
372
|
+
# blazingly fast.
|
373
|
+
strptime = Strptime.new(@time_key_format)
|
374
|
+
Proc.new { |value|
|
375
|
+
value = convert_numeric_time_into_string(value, @time_key_format) if value.is_a?(Numeric)
|
376
|
+
strptime.exec(value).to_datetime
|
377
|
+
}
|
378
|
+
rescue
|
379
|
+
# Can happen if Strptime doesn't recognize the format; or
|
380
|
+
# if strptime couldn't be required (because it's not installed -- it's
|
381
|
+
# ruby 2 only)
|
382
|
+
Proc.new { |value|
|
383
|
+
value = convert_numeric_time_into_string(value, @time_key_format) if value.is_a?(Numeric)
|
384
|
+
DateTime.strptime(value, @time_key_format)
|
385
|
+
}
|
386
|
+
end
|
387
|
+
else
|
388
|
+
Proc.new { |value|
|
389
|
+
value = convert_numeric_time_into_string(value) if value.is_a?(Numeric)
|
390
|
+
DateTime.parse(value)
|
391
|
+
}
|
392
|
+
end
|
393
|
+
end
|
394
|
+
|
395
|
+
def convert_numeric_time_into_string(numeric_time, time_key_format = "%Y-%m-%d %H:%M:%S.%N%z")
|
396
|
+
numeric_time_parser = Fluent::NumericTimeParser.new(:float)
|
397
|
+
Time.at(numeric_time_parser.parse(numeric_time).to_r).strftime(time_key_format)
|
398
|
+
end
|
399
|
+
|
400
|
+
def parse_time(value, event_time, tag)
|
401
|
+
@time_parser.call(value)
|
402
|
+
rescue => e
|
403
|
+
router.emit_error_event(@time_parse_error_tag, Fluent::Engine.now, {'tag' => tag, 'time' => event_time, 'format' => @time_key_format, 'value' => value}, e)
|
404
|
+
return Time.at(event_time).to_datetime
|
405
|
+
end
|
406
|
+
|
407
|
+
def client(host = nil)
|
408
|
+
# check here to see if we already have a client connection for the given host
|
409
|
+
connection_options = get_connection_options(host)
|
410
|
+
|
411
|
+
@_es = nil unless is_existing_connection(connection_options[:hosts])
|
412
|
+
|
413
|
+
@_es ||= begin
|
414
|
+
@current_config = connection_options[:hosts].clone
|
415
|
+
adapter_conf = lambda {|f| f.adapter @http_backend, @backend_options }
|
416
|
+
local_reload_connections = @reload_connections
|
417
|
+
if local_reload_connections && @reload_after > DEFAULT_RELOAD_AFTER
|
418
|
+
local_reload_connections = @reload_after
|
419
|
+
end
|
420
|
+
headers = { 'Content-Type' => @content_type.to_s }.merge(@custom_headers)
|
421
|
+
transport = Elasticsearch::Transport::Transport::HTTP::Faraday.new(connection_options.merge(
|
422
|
+
options: {
|
423
|
+
reload_connections: local_reload_connections,
|
424
|
+
reload_on_failure: @reload_on_failure,
|
425
|
+
resurrect_after: @resurrect_after,
|
426
|
+
logger: @transport_logger,
|
427
|
+
transport_options: {
|
428
|
+
headers: headers,
|
429
|
+
request: { timeout: @request_timeout },
|
430
|
+
ssl: { verify: @ssl_verify, ca_file: @ca_file, version: @ssl_version }
|
431
|
+
},
|
432
|
+
http: {
|
433
|
+
user: @user,
|
434
|
+
password: @password
|
435
|
+
},
|
436
|
+
sniffer_class: @sniffer_class,
|
437
|
+
serializer_class: @serializer_class,
|
438
|
+
}), &adapter_conf)
|
439
|
+
Elasticsearch::Client.new transport: transport
|
440
|
+
end
|
441
|
+
end
|
442
|
+
|
443
|
+
def get_escaped_userinfo(host_str)
|
444
|
+
if m = host_str.match(/(?<scheme>.*)%{(?<user>.*)}:%{(?<password>.*)}(?<path>@.*)/)
|
445
|
+
m["scheme"] +
|
446
|
+
URI.encode_www_form_component(m["user"]) +
|
447
|
+
':' +
|
448
|
+
URI.encode_www_form_component(m["password"]) +
|
449
|
+
m["path"]
|
450
|
+
else
|
451
|
+
host_str
|
452
|
+
end
|
453
|
+
end
|
454
|
+
|
455
|
+
def get_connection_options(con_host=nil)
|
456
|
+
raise "`password` must be present if `user` is present" if @user && !@password
|
457
|
+
|
458
|
+
hosts = if con_host || @hosts
|
459
|
+
(con_host || @hosts).split(',').map do |host_str|
|
460
|
+
# Support legacy hosts format host:port,host:port,host:port...
|
461
|
+
if host_str.match(%r{^[^:]+(\:\d+)?$})
|
462
|
+
{
|
463
|
+
host: host_str.split(':')[0],
|
464
|
+
port: (host_str.split(':')[1] || @port).to_i,
|
465
|
+
scheme: @scheme.to_s
|
466
|
+
}
|
467
|
+
else
|
468
|
+
# New hosts format expects URLs such as http://logs.foo.com,https://john:pass@logs2.foo.com/elastic
|
469
|
+
uri = URI(get_escaped_userinfo(host_str))
|
470
|
+
%w(user password path).inject(host: uri.host, port: uri.port, scheme: uri.scheme) do |hash, key|
|
471
|
+
hash[key.to_sym] = uri.public_send(key) unless uri.public_send(key).nil? || uri.public_send(key) == ''
|
472
|
+
hash
|
473
|
+
end
|
474
|
+
end
|
475
|
+
end.compact
|
476
|
+
else
|
477
|
+
[{host: @host, port: @port, scheme: @scheme.to_s}]
|
478
|
+
end.each do |host|
|
479
|
+
host.merge!(user: @user, password: @password) if !host[:user] && @user
|
480
|
+
host.merge!(path: @path) if !host[:path] && @path
|
481
|
+
end
|
482
|
+
|
483
|
+
{
|
484
|
+
hosts: hosts
|
485
|
+
}
|
486
|
+
end
|
487
|
+
|
488
|
+
def connection_options_description(con_host=nil)
|
489
|
+
get_connection_options(con_host)[:hosts].map do |host_info|
|
490
|
+
attributes = host_info.dup
|
491
|
+
attributes[:password] = 'obfuscated' if attributes.has_key?(:password)
|
492
|
+
attributes.inspect
|
493
|
+
end.join(', ')
|
494
|
+
end
|
495
|
+
|
496
|
+
# append_record_to_messages adds a record to the bulk message
|
497
|
+
# payload to be submitted to Elasticsearch. Records that do
|
498
|
+
# not include '_id' field are skipped when 'write_operation'
|
499
|
+
# is configured for 'create' or 'update'
|
500
|
+
#
|
501
|
+
# returns 'true' if record was appended to the bulk message
|
502
|
+
# and 'false' otherwise
|
503
|
+
def append_record_to_messages(op, meta, header, record, msgs)
|
504
|
+
case op
|
505
|
+
when UPDATE_OP, UPSERT_OP
|
506
|
+
if meta.has_key?(ID_FIELD)
|
507
|
+
header[UPDATE_OP] = meta
|
508
|
+
msgs << @dump_proc.call(header) << BODY_DELIMITER
|
509
|
+
msgs << @dump_proc.call(update_body(record, op)) << BODY_DELIMITER
|
510
|
+
return true
|
511
|
+
end
|
512
|
+
when CREATE_OP
|
513
|
+
if meta.has_key?(ID_FIELD)
|
514
|
+
header[CREATE_OP] = meta
|
515
|
+
msgs << @dump_proc.call(header) << BODY_DELIMITER
|
516
|
+
msgs << @dump_proc.call(record) << BODY_DELIMITER
|
517
|
+
return true
|
518
|
+
end
|
519
|
+
when INDEX_OP
|
520
|
+
header[INDEX_OP] = meta
|
521
|
+
msgs << @dump_proc.call(header) << BODY_DELIMITER
|
522
|
+
msgs << @dump_proc.call(record) << BODY_DELIMITER
|
523
|
+
return true
|
524
|
+
end
|
525
|
+
return false
|
526
|
+
end
|
527
|
+
|
528
|
+
def update_body(record, op)
|
529
|
+
update = remove_keys(record)
|
530
|
+
if @suppress_doc_wrap
|
531
|
+
return update
|
532
|
+
end
|
533
|
+
body = {"doc".freeze => update}
|
534
|
+
if op == UPSERT_OP
|
535
|
+
if update == record
|
536
|
+
body["doc_as_upsert".freeze] = true
|
537
|
+
else
|
538
|
+
body[UPSERT_OP] = record
|
539
|
+
end
|
540
|
+
end
|
541
|
+
body
|
542
|
+
end
|
543
|
+
|
544
|
+
def remove_keys(record)
|
545
|
+
keys = record[@remove_keys_on_update_key] || @remove_keys_on_update || []
|
546
|
+
record.delete(@remove_keys_on_update_key)
|
547
|
+
return record unless keys.any?
|
548
|
+
record = record.dup
|
549
|
+
keys.each { |key| record.delete(key) }
|
550
|
+
record
|
551
|
+
end
|
552
|
+
|
553
|
+
def flatten_record(record, prefix=[])
|
554
|
+
ret = {}
|
555
|
+
if record.is_a? Hash
|
556
|
+
record.each { |key, value|
|
557
|
+
ret.merge! flatten_record(value, prefix + [key.to_s])
|
558
|
+
}
|
559
|
+
elsif record.is_a? Array
|
560
|
+
# Don't mess with arrays, leave them unprocessed
|
561
|
+
ret.merge!({prefix.join(@flatten_hashes_separator) => record})
|
562
|
+
else
|
563
|
+
return {prefix.join(@flatten_hashes_separator) => record}
|
564
|
+
end
|
565
|
+
ret
|
566
|
+
end
|
567
|
+
|
568
|
+
def expand_placeholders(chunk)
|
569
|
+
logstash_prefix = extract_placeholders(@logstash_prefix, chunk)
|
570
|
+
index_name = extract_placeholders(@index_name, chunk)
|
571
|
+
type_name = extract_placeholders(@type_name, chunk)
|
572
|
+
return logstash_prefix, index_name, type_name
|
573
|
+
end
|
574
|
+
|
575
|
+
def multi_workers_ready?
|
576
|
+
true
|
577
|
+
end
|
578
|
+
|
579
|
+
def write(chunk)
|
580
|
+
bulk_message_count = Hash.new { |h,k| h[k] = 0 }
|
581
|
+
bulk_message = Hash.new { |h,k| h[k] = '' }
|
582
|
+
header = {}
|
583
|
+
meta = {}
|
584
|
+
|
585
|
+
tag = chunk.metadata.tag
|
586
|
+
extracted_values = expand_placeholders(chunk)
|
587
|
+
host = if @hosts
|
588
|
+
extract_placeholders(@hosts, chunk)
|
589
|
+
else
|
590
|
+
extract_placeholders(@host, chunk)
|
591
|
+
end
|
592
|
+
|
593
|
+
chunk.msgpack_each do |time, record|
|
594
|
+
next unless record.is_a? Hash
|
595
|
+
begin
|
596
|
+
meta, header, record = process_message(tag, meta, header, time, record, extracted_values)
|
597
|
+
info = if @include_index_in_url
|
598
|
+
RequestInfo.new(host, meta.delete("_index".freeze))
|
599
|
+
else
|
600
|
+
RequestInfo.new(host, nil)
|
601
|
+
end
|
602
|
+
|
603
|
+
if split_request?(bulk_message, info)
|
604
|
+
bulk_message.each do |info, msgs|
|
605
|
+
send_bulk(msgs, tag, chunk, bulk_message_count[info], extracted_values, info) unless msgs.empty?
|
606
|
+
msgs.clear
|
607
|
+
# Clear bulk_message_count for this info.
|
608
|
+
bulk_message_count[info] = 0;
|
609
|
+
next
|
610
|
+
end
|
611
|
+
end
|
612
|
+
|
613
|
+
if append_record_to_messages(@write_operation, meta, header, record, bulk_message[info])
|
614
|
+
bulk_message_count[info] += 1;
|
615
|
+
else
|
616
|
+
if @emit_error_for_missing_id
|
617
|
+
raise MissingIdFieldError, "Missing '_id' field. Write operation is #{@write_operation}"
|
618
|
+
else
|
619
|
+
log.on_debug { log.debug("Dropping record because its missing an '_id' field and write_operation is #{@write_operation}: #{record}") }
|
620
|
+
end
|
621
|
+
end
|
622
|
+
rescue => e
|
623
|
+
router.emit_error_event(tag, time, record, e)
|
624
|
+
end
|
625
|
+
end
|
626
|
+
|
627
|
+
bulk_message.each do |info, msgs|
|
628
|
+
send_bulk(msgs, tag, chunk, bulk_message_count[info], extracted_values, info) unless msgs.empty?
|
629
|
+
msgs.clear
|
630
|
+
end
|
631
|
+
end
|
632
|
+
|
633
|
+
def split_request?(bulk_message, info)
|
634
|
+
# For safety.
|
635
|
+
end
|
636
|
+
|
637
|
+
def split_request_size_check?(bulk_message, info)
|
638
|
+
bulk_message[info].size > @bulk_message_request_threshold
|
639
|
+
end
|
640
|
+
|
641
|
+
def split_request_size_uncheck?(bulk_message, info)
|
642
|
+
false
|
643
|
+
end
|
644
|
+
|
645
|
+
def process_message(tag, meta, header, time, record, extracted_values)
|
646
|
+
logstash_prefix, index_name, type_name = extracted_values
|
647
|
+
|
648
|
+
if @flatten_hashes
|
649
|
+
record = flatten_record(record)
|
650
|
+
end
|
651
|
+
|
652
|
+
dt = nil
|
653
|
+
if @logstash_format || @include_timestamp
|
654
|
+
if record.has_key?(TIMESTAMP_FIELD)
|
655
|
+
rts = record[TIMESTAMP_FIELD]
|
656
|
+
dt = parse_time(rts, time, tag)
|
657
|
+
elsif record.has_key?(@time_key)
|
658
|
+
rts = record[@time_key]
|
659
|
+
dt = parse_time(rts, time, tag)
|
660
|
+
record[TIMESTAMP_FIELD] = dt.iso8601(@time_precision) unless @time_key_exclude_timestamp
|
661
|
+
else
|
662
|
+
dt = Time.at(time).to_datetime
|
663
|
+
record[TIMESTAMP_FIELD] = dt.iso8601(@time_precision)
|
664
|
+
end
|
665
|
+
end
|
666
|
+
|
667
|
+
target_index_parent, target_index_child_key = @target_index_key ? get_parent_of(record, @target_index_key) : nil
|
668
|
+
if target_index_parent && target_index_parent[target_index_child_key]
|
669
|
+
target_index = target_index_parent.delete(target_index_child_key)
|
670
|
+
elsif @logstash_format
|
671
|
+
dt = dt.new_offset(0) if @utc_index
|
672
|
+
target_index = "#{logstash_prefix}#{@logstash_prefix_separator}#{dt.strftime(@logstash_dateformat)}"
|
673
|
+
else
|
674
|
+
target_index = index_name
|
675
|
+
end
|
676
|
+
|
677
|
+
# Change target_index to lower-case since Elasticsearch doesn't
|
678
|
+
# allow upper-case characters in index names.
|
679
|
+
target_index = target_index.downcase
|
680
|
+
if @include_tag_key
|
681
|
+
record[@tag_key] = tag
|
682
|
+
end
|
683
|
+
|
684
|
+
target_type_parent, target_type_child_key = @target_type_key ? get_parent_of(record, @target_type_key) : nil
|
685
|
+
if target_type_parent && target_type_parent[target_type_child_key]
|
686
|
+
target_type = target_type_parent.delete(target_type_child_key)
|
687
|
+
if @last_seen_major_version == 6
|
688
|
+
log.warn "Detected ES 6.x: `@type_name` will be used as the document `_type`."
|
689
|
+
target_type = type_name
|
690
|
+
elsif @last_seen_major_version >= 7
|
691
|
+
log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`."
|
692
|
+
target_type = '_doc'.freeze
|
693
|
+
end
|
694
|
+
else
|
695
|
+
if @last_seen_major_version >= 7 && @type_name != DEFAULT_TYPE_NAME_ES_7x
|
696
|
+
log.warn "Detected ES 7.x or above: `_doc` will be used as the document `_type`."
|
697
|
+
target_type = '_doc'.freeze
|
698
|
+
else
|
699
|
+
target_type = type_name
|
700
|
+
end
|
701
|
+
end
|
702
|
+
|
703
|
+
meta.clear
|
704
|
+
meta["_index".freeze] = target_index
|
705
|
+
meta["_type".freeze] = target_type
|
706
|
+
|
707
|
+
if @pipeline
|
708
|
+
meta["pipeline".freeze] = @pipeline
|
709
|
+
end
|
710
|
+
|
711
|
+
@meta_config_map.each do |record_accessor, meta_key|
|
712
|
+
if raw_value = record_accessor.call(record)
|
713
|
+
meta[meta_key] = raw_value
|
714
|
+
end
|
715
|
+
end
|
716
|
+
|
717
|
+
if @remove_keys
|
718
|
+
@remove_keys.each { |key| record.delete(key) }
|
719
|
+
end
|
720
|
+
|
721
|
+
return [meta, header, record]
|
722
|
+
end
|
723
|
+
|
724
|
+
# returns [parent, child_key] of child described by path array in record's tree
|
725
|
+
# returns [nil, child_key] if path doesnt exist in record
|
726
|
+
def get_parent_of(record, path)
|
727
|
+
parent_object = path[0..-2].reduce(record) { |a, e| a.is_a?(Hash) ? a[e] : nil }
|
728
|
+
[parent_object, path[-1]]
|
729
|
+
end
|
730
|
+
|
731
|
+
# send_bulk given a specific bulk request, the original tag,
|
732
|
+
# chunk, and bulk_message_count
|
733
|
+
def send_bulk(data, tag, chunk, bulk_message_count, extracted_values, info)
|
734
|
+
begin
|
735
|
+
|
736
|
+
log.on_trace { log.trace "bulk request: #{data}" }
|
737
|
+
response = client(info.host).bulk body: data, index: info.index
|
738
|
+
log.on_trace { log.trace "bulk response: #{response}" }
|
739
|
+
|
740
|
+
if response['errors']
|
741
|
+
error = Fluent::Plugin::ElasticsearchErrorHandler.new(self)
|
742
|
+
error.handle_error(response, tag, chunk, bulk_message_count, extracted_values)
|
743
|
+
end
|
744
|
+
rescue RetryStreamError => e
|
745
|
+
emit_tag = @retry_tag ? @retry_tag : tag
|
746
|
+
router.emit_stream(emit_tag, e.retry_stream)
|
747
|
+
rescue => e
|
748
|
+
ignore = @ignore_exception_classes.any? { |clazz| e.class <= clazz }
|
749
|
+
|
750
|
+
log.warn "Exception ignored in tag #{tag}: #{e.class.name} #{e.message}" if ignore
|
751
|
+
|
752
|
+
@_es = nil if @reconnect_on_error
|
753
|
+
@_es_info = nil if @reconnect_on_error
|
754
|
+
|
755
|
+
raise UnrecoverableRequestFailure if ignore && @exception_backup
|
756
|
+
|
757
|
+
# FIXME: identify unrecoverable errors and raise UnrecoverableRequestFailure instead
|
758
|
+
raise RecoverableRequestFailure, "could not push logs to Elasticsearch cluster (#{connection_options_description(info.host)}): #{e.message}" unless ignore
|
759
|
+
end
|
760
|
+
end
|
761
|
+
|
762
|
+
def is_existing_connection(host)
|
763
|
+
# check if the host provided match the current connection
|
764
|
+
return false if @_es.nil?
|
765
|
+
return false if @current_config.nil?
|
766
|
+
return false if host.length != @current_config.length
|
767
|
+
|
768
|
+
for i in 0...host.length
|
769
|
+
if !host[i][:host].eql? @current_config[i][:host] || host[i][:port] != @current_config[i][:port]
|
770
|
+
return false
|
771
|
+
end
|
772
|
+
end
|
773
|
+
|
774
|
+
return true
|
775
|
+
end
|
776
|
+
end
|
777
|
+
end
|