Class: Fluent::Plugin::PrometheusOutputMonitorInput

Inherits:
Input
  • Object
show all
Includes:
PrometheusLabelParser
Defined in:
lib/fluent/plugin/in_prometheus_output_monitor.rb

Constant Summary collapse

MONITOR_IVARS = 
[
  :retry,

  :num_errors,
  :emit_count,

  # for v0.12
  :last_retry_time,

  # from v0.14
  :emit_records,
  :write_count,
  :rollback_count,

  # from v1.6.0
  :flush_time_count,
  :slow_flush_count,
]

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from PrometheusLabelParser

#parse_labels_elements

Constructor Details

#initialize ⇒ PrometheusOutputMonitorInput

Returns a new instance of PrometheusOutputMonitorInput.



35
36
37
38
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 35

def initialize
  super
  @registry = ::Prometheus::Client.registry
end

Instance Attribute Details

#registry ⇒ Object (readonly)

Returns the value of attribute registry.



14
15
16
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 14

def registry
  @registry
end

Instance Method Details

#configure(conf) ⇒ Object 



44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 44

def configure(conf)
  super
  hostname = Socket.gethostname
  expander_builder = Fluent::Plugin::Prometheus.placeholder_expander(log)
  expander = expander_builder.build({ 'hostname' => hostname, 'worker_id' => fluentd_worker_id })
  @base_labels = parse_labels_elements(conf)
  @base_labels.each do |key, value|
    unless value.is_a?(String)
      raise Fluent::ConfigError, "record accessor syntax is not available in prometheus_output_monitor"
    end
    @base_labels[key] = expander.expand(value)
  end

  @monitor_agent = Fluent::Plugin::MonitorAgentInput.new

  @gauge_or_counter = @gauge_all ? :gauge : :counter
end

#get_gauge(name, docstring) ⇒ Object 



217
218
219
220
221
222
223
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 217

def get_gauge(name, docstring)
  if @registry.exist?(name)
    @registry.get(name)
  else
    @registry.gauge(name, docstring: docstring, labels: @base_labels.keys + [:plugin_id, :type])
  end
end

#get_gauge_or_counter(name, docstring) ⇒ Object 



225
226
227
228
229
230
231
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 225

def get_gauge_or_counter(name, docstring)
  if @registry.exist?(name)
    @registry.get(name)
  else
    @registry.public_send(@gauge_or_counter, name, docstring: docstring, labels: @base_labels.keys + [:plugin_id, :type])
  end
end

#labels(plugin_info) ⇒ Object 



210
211
212
213
214
215
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 210

def labels(plugin_info)
  @base_labels.merge(
    plugin_id: plugin_info["plugin_id"],
    type: plugin_info["type"],
  )
end

#multi_workers_ready? ⇒ Boolean

Returns:

  • (Boolean) 


40
41
42
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 40

def multi_workers_ready?
  true
end

#start ⇒ Object 



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
90
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
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 62

def start
  super

  @metrics = {
    # Buffer metrics
    buffer_total_queued_size: get_gauge(
      :fluentd_output_status_buffer_total_bytes,
      'Current total size of stage and queue buffers.'),
    buffer_stage_length: get_gauge(
      :fluentd_output_status_buffer_stage_length,
      'Current length of stage buffers.'),
    buffer_stage_byte_size: get_gauge(
      :fluentd_output_status_buffer_stage_byte_size,
      'Current total size of stage buffers.'),
    buffer_queue_length: get_gauge(
      :fluentd_output_status_buffer_queue_length,
      'Current length of queue buffers.'),
    buffer_queue_byte_size: get_gauge(
      :fluentd_output_status_queue_byte_size,
      'Current total size of queue buffers.'),
    buffer_available_buffer_space_ratios: get_gauge(
      :fluentd_output_status_buffer_available_space_ratio,
      'Ratio of available space in buffer.'),
    buffer_newest_timekey: get_gauge(
      :fluentd_output_status_buffer_newest_timekey,
      'Newest timekey in buffer.'),
    buffer_oldest_timekey: get_gauge(
      :fluentd_output_status_buffer_oldest_timekey,
      'Oldest timekey in buffer.'),

    # Output metrics
    retry_counts: get_gauge_or_counter(
      :fluentd_output_status_retry_count,
      'Current retry counts.'),
    num_errors: get_gauge_or_counter(
      :fluentd_output_status_num_errors,
      'Current number of errors.'),
    emit_count: get_gauge_or_counter(
      :fluentd_output_status_emit_count,
      'Current emit counts.'),
    emit_records: get_gauge_or_counter(
      :fluentd_output_status_emit_records,
      'Current emit records.'),
    write_count: get_gauge_or_counter(
      :fluentd_output_status_write_count,
      'Current write counts.'),
    rollback_count: get_gauge(
      :fluentd_output_status_rollback_count,
      'Current rollback counts.'),
    flush_time_count: get_gauge_or_counter(
      :fluentd_output_status_flush_time_count,
      'Total flush time.'),
    slow_flush_count: get_gauge_or_counter(
      :fluentd_output_status_slow_flush_count,
      'Current slow flush counts.'),
    retry_wait: get_gauge(
      :fluentd_output_status_retry_wait,
      'Current retry wait'),
  }
  timer_execute(:in_prometheus_output_monitor, @interval, &method(:update_monitor_info))
end

#update_monitor_info ⇒ Object 



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
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
 
# File 'lib/fluent/plugin/in_prometheus_output_monitor.rb', line 124

def update_monitor_info
  opts = {
    ivars: MONITOR_IVARS,
    with_retry: true,
  }

  agent_info = @monitor_agent.plugins_info_all(opts).select {|info|
    info['plugin_category'] == 'output'.freeze
  }

  monitor_info = {
    # buffer metrics
    'buffer_total_queued_size' => @metrics[:buffer_total_queued_size],
    'buffer_stage_length' => @metrics[:buffer_stage_length],
    'buffer_stage_byte_size' => @metrics[:buffer_stage_byte_size],
    'buffer_queue_length' => @metrics[:buffer_queue_length],
    'buffer_queue_byte_size' => @metrics[:buffer_queue_byte_size],
    'buffer_available_buffer_space_ratios' => @metrics[:buffer_available_buffer_space_ratios],
    'buffer_newest_timekey' => @metrics[:buffer_newest_timekey],
    'buffer_oldest_timekey' => @metrics[:buffer_oldest_timekey],

    # output metrics
    'retry_count' => @metrics[:retry_counts],
    # Needed since Fluentd v1.14 due to metrics extensions.
    'num_errors' => @metrics[:num_errors],
    'write_count' => @metrics[:write_count],
    'emit_count' => @metrics[:emit_count],
    'emit_records' => @metrics[:emit_records],
    'rollback_count' => @metrics[:rollback_count],
    'flush_time_count' => @metrics[:flush_time_count],
    'slow_flush_count' => @metrics[:slow_flush_count],
  }
  # No needed for Fluentd v1.14 but leave as-is for backward compatibility.
  instance_vars_info = {
    num_errors: @metrics[:num_errors],
    write_count: @metrics[:write_count],
    emit_count: @metrics[:emit_count],
    emit_records: @metrics[:emit_records],
    rollback_count: @metrics[:rollback_count],
    flush_time_count: @metrics[:flush_time_count],
    slow_flush_count: @metrics[:slow_flush_count],
  }

  agent_info.each do |info|
    label = labels(info)

    monitor_info.each do |name, metric|
      if info[name]
        if metric.is_a?(::Prometheus::Client::Gauge)
          metric.set(info[name], labels: label)
        elsif metric.is_a?(::Prometheus::Client::Counter)
          metric.increment(by: info[name] - metric.get(labels: label), labels: label)
        end
      end
    end

    if info['instance_variables']
      instance_vars_info.each do |name, metric|
        if info['instance_variables'][name]
          if metric.is_a?(::Prometheus::Client::Gauge)
            metric.set(info['instance_variables'][name], labels: label)
          elsif metric.is_a?(::Prometheus::Client::Counter)
            metric.increment(by: info['instance_variables'][name] - metric.get(labels: label), labels: label)
          end
        end
      end
    end

    # compute current retry_wait
    if info['retry']
      next_time = info['retry']['next_time']
      start_time = info['retry']['start']
      if start_time.nil? && info['instance_variables']
        # v0.12 does not include start, use last_retry_time instead
        start_time = info['instance_variables'][:last_retry_time]
      end

      wait = 0
      if next_time && start_time
        wait = next_time - start_time
      end
      @metrics[:retry_wait].set(wait.to_f, labels: label)
    end
  end
end