Class: Plugin::InfluxdbDeduplicationFilter

Inherits:
Filter
  • Object
show all
Defined in:
lib/fluent/plugin/filter_influxdb_deduplication.rb

Instance Method Summary collapse

Instance Method Details

#configure(conf) ⇒ Object 



20
21
22
23
24
25
26
27
28
29
30
31
32
 
# File 'lib/fluent/plugin/filter_influxdb_deduplication.rb', line 20

def configure(conf)
  super

  if @time == nil and @tag == nil
    raise Fluent::ConfigError, "one of tag or time deduplication needs to be set."
  elsif @time != nil and @tag != nil
    raise Fluent::ConfigError, "tag and time deduplication are mutually exclusive."
  elsif @time != nil and (@time.key == nil or @time.key == "")
    raise Fluent::ConfigError, "an output 'key' field is required for time deduplication"
  elsif @tag != nil and (@tag == nil or @tag.key == "")
    raise Fluent::ConfigError, "an output 'key' field is required for tag deduplication"
  end
end

#filter(tag, time, record) ⇒ Object 



41
42
43
44
45
46
47
 
# File 'lib/fluent/plugin/filter_influxdb_deduplication.rb', line 41

def filter(tag, time, record)
  if @time
    time_deduplication(time, record)
  else
    tag_deduplication(time, record)
  end
end

#start ⇒ Object 



34
35
36
37
38
39
 
# File 'lib/fluent/plugin/filter_influxdb_deduplication.rb', line 34

def start
  super

  @last_timestamp = 0
  @sequence = 0
end

#tag_deduplication(time, record) ⇒ Object 



91
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
120
121
122
123
124
125
 
# File 'lib/fluent/plugin/filter_influxdb_deduplication.rb', line 91

def tag_deduplication(time, record)
  if time.is_a?(Integer)
    input_time = time
  elsif time.is_a?(Fluent::EventTime)
    input_time = time.sec * 1000000000 + time.nsec
  else
    @log.error("unreadable time")
    return nil
  end

  if input_time < @last_timestamp
    @log.debug("out of sequence timestamp")
    if @order_key
      record[@order_key] = false
    else
      @log.debug("out of order record dropped")
      return nil
    end
  elsif input_time == @last_timestamp
    @sequence = @sequence + 1
    record[@tag.key] = @sequence
    if @order_key
      record[@order_key] = true
    end
  else
    @sequence = 0
    @last_timestamp = input_time
    record[@tag.key] = 0
    if @order_key
      record[@order_key] = true
    end
  end

  record
end

#time_deduplication(time, record) ⇒ Object 



49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
 
# File 'lib/fluent/plugin/filter_influxdb_deduplication.rb', line 49

def time_deduplication(time, record)
  if time.is_a?(Integer)
    input_time = Fluent::EventTime.new(time)
  elsif time.is_a?(Fluent::EventTime)
    input_time = time
  else
    @log.error("unreadable time")
    return nil
  end

  nano_time = input_time.sec * 1000000000

  if input_time.sec < @last_timestamp
    @log.debug("out of sequence timestamp")
    if @order_key
      record[@order_key] = false
      record[@time.key] = nano_time
    else
      @log.debug("out of order record dropped")
      return nil
    end
  elsif input_time.sec == @last_timestamp and @sequence < 999999999
    @sequence = @sequence + 1
    record[@time.key] = nano_time + @sequence
    if @order_key
      record[@order_key] = true
    end
  elsif input_time.sec == @last_timestamp and @sequence == 999999999
    @log.error("received more then 999999999 records in a second")
    return nil
  else
    @sequence = 0
    @last_timestamp = input_time.sec
    record[@time.key] = nano_time
    if @order_key
      record[@order_key] = true
    end
  end

  record
end