Class: OpenC3::ReducerMicroservice

Inherits:
Microservice show all
Defined in:
lib/openc3/microservices/reducer_microservice.rb

Constant Summary collapse

MINUTE_METRIC = 
'reducer_minute_processing_seconds'
HOUR_METRIC = 
'reducer_hour_processing_seconds'
DAY_METRIC = 
'reducer_day_processing_seconds'
SHUTDOWN_DELAY_SECS =

How long to wait for any currently running jobs to complete before killing them

5
MINUTE_ENTRY_NSECS = 
60 * 1_000_000_000
MINUTE_FILE_NSECS = 
3600 * 1_000_000_000
HOUR_ENTRY_NSECS = 
3600 * 1_000_000_000
HOUR_FILE_NSECS = 
3600 * 24 * 1_000_000_000
DAY_ENTRY_NSECS = 
3600 * 24 * 1_000_000_000
DAY_FILE_NSECS = 
3600 * 24 * 5 * 1_000_000_000

Instance Attribute Summary

Attributes inherited from Microservice

#count, #custom, #error, #logger, #microservice_status_thread, #name, #scope, #secrets, #state

Instance Method Summary collapse

Methods inherited from Microservice

#as_json, run

Constructor Details

#initialize(name) ⇒ ReducerMicroservice

Returns a new instance of ReducerMicroservice.

Parameters:

  • name (String) —

    Microservice name formatted as <SCOPE>__REDUCER__<TARGET> where <SCOPE> and <TARGET> are variables representing the scope name and target name



91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 91

def initialize(name)
  super(name, is_plugin: false)

  if @config['options']
    @config['options'].each do |option|
      case option[0].upcase
      when 'BUFFER_DEPTH' # Buffer depth to write in time order
        @buffer_depth = option[1].to_i
      else
        @logger.error("Unknown option passed to microservice #{@name}: #{option}")
      end
    end
  end

  @buffer_depth = 10 unless @buffer_depth
  @target_name = name.split('__')[-1]
  @packet_logs = {}

  @error_count = 0
  @metric.set(name: 'reducer_total', value: @count, type: 'counter')
  @metric.set(name: 'reducer_error_total', value: @error_count, type: 'counter')
end

Instance Method Details

#check_new_file(reducer_state, plw, type, target_name, stored, current_time, file_nanoseconds) ⇒ Object 



383
384
385
386
387
388
389
390
391
392
393
394
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 383

def check_new_file(reducer_state, plw, type, target_name, stored, current_time, file_nanoseconds)
  plw_first_time_nsec = plw.buffered_first_time_nsec
  if plw_first_time_nsec && ((current_time - plw_first_time_nsec) >= file_nanoseconds)
    # Write out all entries in progress
    write_all_entries(reducer_state, plw, type, target_name, stored)
    reducer_state.clear
    plw.close_file
    return true
  else
    return false
  end
end

#extract_entry_samples(packet) ⇒ Object

Extract just the not reduced fields from a JsonPacket



514
515
516
517
518
519
520
521
522
523
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 514

def extract_entry_samples(packet)
  result = {}
  packet.json_hash.each do |key, value|
    key_split = key.split('__')
    if (not key_split[1] or not ['N', 'X', 'A', 'S'].include?(key_split[1][-1])) and key != '_NUM_SAMPLES'
      result[key] = value
    end
  end
  return result
end

#metric(name) ⇒ Object 



151
152
153
154
155
156
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 151

def metric(name)
  start = Process.clock_gettime(Process::CLOCK_MONOTONIC)
  yield
  elapsed = Process.clock_gettime(Process::CLOCK_MONOTONIC) - start # seconds as a float
  @metric.set(name: name, value: elapsed, type: 'gauge', unit: 'seconds')
end

#process_file(filename, type, entry_nanoseconds, file_nanoseconds) ⇒ Object 



191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
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
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 191

def process_file(filename, type, entry_nanoseconds, file_nanoseconds)
  file = BucketFile.new(filename)
  file.retrieve
  unless file.local_path
    @logger.warn("Reducer Warning: #{filename}: Could not be retrieved")
    return
  end

  # Determine if we already have a PacketLogWriter created
  _, _, scope, target_name, _, rt_or_stored, _ = File.basename(filename).split('__')
  stored = (rt_or_stored == "stored")

  if @target_name != target_name
    raise "Target name in file #{filename} does not match microservice target name #{@target_name}"
  end
  plw = @packet_logs["#{scope}__#{target_name}__#{rt_or_stored}__#{type}"]
  unless plw
    # Create a new PacketLogWriter for this reduced data
    # e.g. DEFAULT/reduced_minute_logs/tlm/INST/20220101/
    # 20220101204857274290500__20220101205857276524900__DEFAULT__INST__ALL__rt__reduced_minute.bin
    remote_log_directory = "#{scope}/reduced_#{type}_logs/tlm/#{target_name}"
    label = "#{scope}__#{target_name}__ALL__#{rt_or_stored}__reduced_#{type}"
    plw = BufferedPacketLogWriter.new(remote_log_directory, label, true, nil, 1_000_000_000, nil, nil, true, @buffer_depth)
    @packet_logs["#{scope}__#{target_name}__#{rt_or_stored}__#{type}"] = plw
  end

  # The lifetime of all these variables is a single file - single target / multiple packets
  reducer_state = {}
  plr = OpenC3::PacketLogReader.new
  plr.each(file.local_path) do |packet|
    # Check to see if we should start a new log file before processing this packet
    current_time = packet.packet_time.to_nsec_from_epoch
    check_new_file(reducer_state, plw, type, target_name, stored, current_time, file_nanoseconds)
    state = setup_state(reducer_state, packet, current_time)

    # Determine if we've rolled over a entry boundary
    # We have to use current % entry_nanoseconds < previous % entry_nanoseconds because
    # we don't know the data rates. We also have to check for current - previous >= entry_nanoseconds
    # in case the data rate is so slow we don't have multiple samples per entry
    if state.previous_time &&
      (
        (state.current_time % entry_nanoseconds < state.previous_time % entry_nanoseconds) || # Try to create at perfect intervals
          (state.current_time - state.previous_time >= entry_nanoseconds)                 # Handle big gaps
      )
      write_entry(state, plw, type, target_name, packet.packet_name, stored)
      if check_new_file(reducer_state, plw, type, target_name, stored, current_time, file_nanoseconds)
        state = setup_state(reducer_state, packet, current_time)
      end
    end

    if type == 'minute'
      state.entry_samples ||= packet.json_hash.dup # Grab all the samples from the first packet
      if state.first
        state.raw_values = packet.read_all(:RAW, nil, packet.read_all_names(:RAW)).select { |key, value| value.is_a?(Numeric) }
        state.raw_keys ||= state.raw_values.keys
        state.converted_values = packet.read_all(:CONVERTED, nil, packet.read_all_names(:CONVERTED)).select { |key, value| value.is_a?(Numeric) }
        state.converted_keys ||= state.converted_values.keys
      else
        state.raw_values = packet.read_all(:RAW, nil, state.raw_keys).select { |key, value| value.is_a?(Numeric) }
        state.converted_values = packet.read_all(:CONVERTED, nil, state.converted_keys).select { |key, value| value.is_a?(Numeric) }
      end
    else
      # Hour or Day
      state.entry_samples ||= extract_entry_samples(packet)
      if state.first
        state.raw_max_values = packet.read_all(:RAW, :MAX, packet.read_all_names(:RAW, :MAX))
        state.raw_keys = state.raw_max_values.keys
        state.converted_max_values = packet.read_all(:CONVERTED, :MAX, packet.read_all_names(:CONVERTED, :MAX))
        state.converted_keys = state.converted_max_values.keys
      else
        state.raw_max_values = packet.read_all(:RAW, :MAX, state.raw_keys)
        state.converted_max_values = packet.read_all(:CONVERTED, :MAX, state.converted_keys)
      end
      state.raw_min_values = packet.read_all(:RAW, :MIN, state.raw_keys)
      state.raw_avg_values = packet.read_all(:RAW, :AVG, state.raw_keys)
      state.raw_stddev_values = packet.read_all(:RAW, :STDDEV, state.raw_keys)
      state.converted_min_values = packet.read_all(:CONVERTED, :MIN, state.converted_keys)
      state.converted_avg_values = packet.read_all(:CONVERTED, :AVG, state.converted_keys)
      state.converted_stddev_values = packet.read_all(:CONVERTED, :STDDEV, state.converted_keys)
    end

    reduced = state.reduced
    if type == 'minute'
      # Update statistics for this packet's raw values
      state.raw_values.each do |key, value|
        if value
          reduced["#{key}__VALS"] ||= []
          reduced["#{key}__VALS"] << value
          reduced["#{key}__N"] ||= value
          reduced["#{key}__N"] = value if value < reduced["#{key}__N"]
          reduced["#{key}__X"] ||= value
          reduced["#{key}__X"] = value if value > reduced["#{key}__X"]
        end
      end

      # Update statistics for this packet's converted values
      state.converted_values.each do |key, value|
        if value
          reduced["#{key}__CVALS"] ||= []
          reduced["#{key}__CVALS"] << value
          reduced["#{key}__CN"] ||= value
          reduced["#{key}__CN"] = value if value < reduced["#{key}__CN"]
          reduced["#{key}__CX"] ||= value
          reduced["#{key}__CX"] = value if value > reduced["#{key}__CX"]
        end
      end
    else
      # Update statistics for this packet's raw values
      state.raw_max_values.each do |key, value|
        if value
          max_key = "#{key}__X"
          reduced[max_key] ||= value
          reduced[max_key] = value if value > reduced[max_key]
        end
      end
      state.raw_min_values.each do |key, value|
        if value
          min_key = "#{key}__N"
          reduced[min_key] ||= value
          reduced[min_key] = value if value < reduced[min_key]
        end
      end
      state.raw_avg_values.each do |key, value|
        if value
          avg_values_key = "#{key}__AVGVALS"
          reduced[avg_values_key] ||= []
          reduced[avg_values_key] << value
        end
      end
      state.raw_stddev_values.each do |key, value|
        if value
          stddev_values_key = "#{key}__STDDEVVALS"
          reduced[stddev_values_key] ||= []
          reduced[stddev_values_key] << value
        end
      end

      # Update statistics for this packet's converted values
      state.converted_max_values.each do |key, value|
        if value
          max_key = "#{key}__CX"
          reduced[max_key] ||= value
          reduced[max_key] = value if value > reduced[max_key]
        end
      end
      state.converted_min_values.each do |key, value|
        if value
          min_key = "#{key}__CN"
          reduced[min_key] ||= value
          reduced[min_key] = value if value < reduced[min_key]
        end
      end
      state.converted_avg_values.each do |key, value|
        if value
          avg_values_key = "#{key}__CAVGVALS"
          reduced[avg_values_key] ||= []
          reduced[avg_values_key] << value
        end
      end
      state.converted_stddev_values.each do |key, value|
        if value
          stddev_values_key = "#{key}__CSTDDEVVALS"
          reduced[stddev_values_key] ||= []
          reduced[stddev_values_key] << value
        end
      end

      reduced["_NUM_SAMPLES__VALS"] ||= []
      reduced["_NUM_SAMPLES__VALS"] << packet.read('_NUM_SAMPLES')
    end

    state.first = false
  end
  file.delete # Remove the local copy

  write_all_entries(reducer_state, plw, type, target_name, stored)

  @count += 1
  @metric.set(name: 'reducer_total', value: @count, type: 'counter')

  true
rescue => e
  if file.local_path and File.exist?(file.local_path)
    @logger.error("Reducer Error: #{filename}: #{File.size(file.local_path)} bytes: \n#{e.formatted}")
  else
    @logger.error("Reducer Error: #{filename}: \n#{e.formatted}")
  end
  @error_count += 1
  @metric.set(name: 'reducer_error_total', value: @error_count, type: 'counter')
  false
end

#reduce(type, raw_keys, converted_keys, reduced) ⇒ Object 



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
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 449

def reduce(type, raw_keys, converted_keys, reduced)
  # We've collected all the values so calculate the AVG and STDDEV
  if type == 'minute'
    raw_keys.each do |key|
      if reduced["#{key}__VALS"]
        reduced["_NUM_SAMPLES"] ||= reduced["#{key}__VALS"].length # Keep a single sample count per packet
        reduced["#{key}__A"], reduced["#{key}__S"] =
          Math.stddev_population(reduced["#{key}__VALS"])
        # Remove the raw values as they're only used for AVG / STDDEV calculation
        reduced.delete("#{key}__VALS")
      end
    end

    converted_keys.each do |key|
      if reduced["#{key}__CVALS"]
        reduced["_NUM_SAMPLES"] ||= reduced["#{key}__CVALS"].length # Keep a single sample count per packet
        reduced["#{key}__CA"], reduced["#{key}__CS"] =
          Math.stddev_population(reduced["#{key}__CVALS"])

        # Remove the converted values as they're only used for AVG / STDDEV calculation
        reduced.delete("#{key}__CVALS")
      end
    end
  else
    samples = reduced["_NUM_SAMPLES__VALS"]
    samples_sum = samples.sum
    reduced["_NUM_SAMPLES"] = samples_sum
    reduced.delete("_NUM_SAMPLES__VALS")

    raw_keys.each { |key| reduce_running(key, reduced, samples, samples_sum, "__A", "__S", "__AVGVALS", "__STDDEVVALS") }
    converted_keys.each { |key| reduce_running(key, reduced, samples, samples_sum, "__CA", "__CS", "__CAVGVALS", "__CSTDDEVVALS") }
  end
end

#reduce_day ⇒ Object 



180
181
182
183
184
185
186
187
188
189
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 180

def reduce_day
  metric(DAY_METRIC) do
    ReducerModel
      .all_files(type: :HOUR, target: @target_name, scope: @scope)
      .each do |file|
        process_file(file, 'day', DAY_ENTRY_NSECS, DAY_FILE_NSECS)
        ReducerModel.rm_file(file)
      end
  end
end

#reduce_hour ⇒ Object 



169
170
171
172
173
174
175
176
177
178
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 169

def reduce_hour
  metric(HOUR_METRIC) do
    ReducerModel
      .all_files(type: :MINUTE, target: @target_name, scope: @scope)
      .each do |file|
        process_file(file, 'hour', HOUR_ENTRY_NSECS, HOUR_FILE_NSECS)
        ReducerModel.rm_file(file)
      end
  end
end

#reduce_minute ⇒ Object 



158
159
160
161
162
163
164
165
166
167
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 158

def reduce_minute
  metric(MINUTE_METRIC) do
    ReducerModel
      .all_files(type: :DECOM, target: @target_name, scope: @scope)
      .each do |file|
        process_file(file, 'minute', MINUTE_ENTRY_NSECS, MINUTE_FILE_NSECS)
        ReducerModel.rm_file(file)
      end
  end
end

#reduce_running(key, reduced, samples, samples_sum, avg_key, stddev_key, avgvals_key, stddevvals_key) ⇒ Object 



483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 483

def reduce_running(key, reduced, samples, samples_sum, avg_key, stddev_key, avgvals_key, stddevvals_key)
  # Calculate Average
  weighted_sum = 0
  avg = reduced["#{key}#{avgvals_key}"]
  avg.each_with_index do |val, i|
    weighted_sum += (val * samples[i])
  end
  reduced["#{key}#{avg_key}"] = weighted_sum / samples_sum

  # Do the STDDEV calc last so we can use the previously calculated AVG
  # See https://math.stackexchange.com/questions/1547141/aggregating-standard-deviation-to-a-summary-point
  s2 = 0
  reduced["#{key}#{stddevvals_key}"].each_with_index do |val, i|
    # puts "i:#{i} val:#{val} samples[i]:#{samples[i]} avg[i]:#{avg[i]}"
    s2 += (samples[i] * avg[i]**2 + val**2)
  end

  # Note: For very large numbers with very small deviations this sqrt can fail.
  # If so then just set the stddev to 0.
  begin
    reduced["#{key}#{stddev_key}"] =
      Math.sqrt(s2 / samples_sum - reduced["#{key}#{avg_key}"])
  rescue Exception
    reduced["#{key}#{stddev_key}"] = 0.0
  end

  reduced.delete("#{key}#{avgvals_key}")
  reduced.delete("#{key}#{stddevvals_key}")
end

#run ⇒ Object 



114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 114

def run
  # Note it takes several seconds to create the scheduler
  @scheduler = Rufus::Scheduler.new
  # Run every minute
  @scheduler.cron '* * * * *', first: :now do
    reduce_minute
  end
  # Run every 15 minutes
  @scheduler.cron '*/15 * * * *', first: :now do
    reduce_hour
  end
  # Run hourly at minute 5 to allow the hour reducer to finish
  @scheduler.cron '5 * * * *', first: :now do
    reduce_day
  end

  # Let the current thread join the scheduler thread and
  # block until shutdown is called
  @scheduler.join
end

#setup_state(reducer_state, packet, current_time) ⇒ Object 



396
397
398
399
400
401
402
403
404
405
406
407
408
409
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 396

def setup_state(reducer_state, packet, current_time)
  # Get state for this packet
  state = reducer_state[packet.packet_name]
  unless state
    state = ReducerState.new
    reducer_state[packet.packet_name] = state
  end

  # Update state timestamps
  state.previous_time = state.current_time # Will be nil first packet
  state.current_time = current_time
  state.entry_time ||= state.current_time # Sets the entry time from the first packet
  return state
end

#shutdown ⇒ Object 



135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 135

def shutdown
  @logger.info("Shutting down reducer microservice: #{@name}")
  @scheduler.shutdown(wait: SHUTDOWN_DELAY_SECS) if @scheduler

  # Make sure all the existing logs are properly closed down
  threads = []
  @packet_logs.each do |name, log|
    threads.concat(log.shutdown)
  end
  # Wait for all the logging threads to move files to buckets
  threads.flatten.compact.each do |thread|
    thread.join
  end
  super()
end

#write_all_entries(reducer_state, plw, type, target_name, stored) ⇒ Object 



411
412
413
414
415
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 411

def write_all_entries(reducer_state, plw, type, target_name, stored)
  reducer_state.each do |packet_name, state|
    write_entry(state, plw, type, target_name, packet_name, stored)
  end
end

#write_entry(state, plw, type, target_name, packet_name, stored) ⇒ Object 



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
 
# File 'lib/openc3/microservices/reducer_microservice.rb', line 417

def write_entry(state, plw, type, target_name, packet_name, stored)
  return unless state.reduced.length > 0
  reduce(type, state.raw_keys, state.converted_keys, state.reduced)
  state.reduced.merge!(state.entry_samples)
  time = state.entry_time
  data = JSON.generate(state.reduced.as_json(:allow_nan => true))
  if type == "minute"
    redis_topic, redis_offset = TelemetryReducedMinuteTopic.write(target_name: target_name, packet_name: packet_name, stored: stored, time: time, data: data, scope: @scope)
  elsif type == "hour"
    redis_topic, redis_offset = TelemetryReducedHourTopic.write(target_name: target_name, packet_name: packet_name, stored: stored, time: time, data: data, scope: @scope)
  else
    redis_topic, redis_offset = TelemetryReducedDayTopic.write(target_name: target_name, packet_name: packet_name, stored: stored, time: time, data: data, scope: @scope)
  end
  plw.buffered_write(
    :JSON_PACKET,
    :TLM,
    target_name,
    packet_name,
    time,
    stored,
    data,
    nil,
    redis_topic,
    redis_offset
  )

  # Reset necessary state variables
  state.entry_time = state.current_time # This packet starts the next entry
  state.entry_samples = nil
  state.reduced = {}
end