Class: DeadBro::MemoryTrackingSubscriber
- Inherits:
-
Object
- Object
- DeadBro::MemoryTrackingSubscriber
- Defined in:
- lib/dead_bro/memory_tracking_subscriber.rb
Constant Summary collapse
- PROCESS_ACTION_EVENT =
Allocation counts come from the process_action event (Rails instruments allocations there via ActiveSupport::Notifications). The old “object_allocations.active_support” constant was never emitted by Rails, so that subscription was dead code — removed.
"process_action.action_controller"- THREAD_LOCAL_KEY =
:dead_bro_memory_events- LARGE_OBJECT_THRESHOLD =
Consider objects larger than this many bytes as “large”
1_000_000- MAX_ALLOCATIONS_PER_REQUEST =
Performance optimization settings
1000- LARGE_OBJECT_SAMPLE_RATE =
Limit allocations tracked per request
0.01- MAX_LARGE_OBJECTS =
Sample 1% of live objects to estimate large ones
50
Class Method Summary collapse
- .analyze_large_objects(large_objects) ⇒ Object
- .analyze_memory_performance(memory_events) ⇒ Object
- .analyze_memory_trends(snapshots) ⇒ Object
- .calculate_gc_efficiency(gc_before, gc_after) ⇒ Object
- .count_objects_snapshot ⇒ Object
-
.current_events ⇒ Object
Current frame (top of stack) for nested job tracking; nil if none.
- .gc_stats ⇒ Object
- .heap_pages ⇒ Object
- .memory_usage_mb ⇒ Object
- .object_count ⇒ Object
- .object_space_available? ⇒ Boolean
-
.record_request_allocations(allocations:, allocated_bytes:) ⇒ Object
Record request-level allocation counters from Rails instrumentation.
- .sample_large_objects ⇒ Object
-
.snapshot_at(label) ⇒ Object
Helper method to take memory snapshots at specific points.
- .start_request_tracking ⇒ Object
- .stop_request_tracking ⇒ Object
-
.subscribe!(client: Client.new) ⇒ Object
Cap number of large objects captured per request.
- .take_memory_snapshot(label = nil) ⇒ Object
- .top_object_type_deltas(deltas, limit: 10) ⇒ Object
- .track_allocation(data, started, finished) ⇒ Object
Class Method Details
.analyze_large_objects(large_objects) ⇒ Object
241 242 243 244 245 246 247 248 249 250 251 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 241 def self.analyze_large_objects(large_objects) return {} if large_objects.empty? { count: large_objects.count, total_size_mb: large_objects.sum { |obj| obj[:size_mb] }.round(2), largest_object_mb: large_objects.max_by { |obj| obj[:size_mb] }[:size_mb], by_class: large_objects.group_by { |obj| obj[:class_name] } .transform_values(&:count) } end |
.analyze_memory_performance(memory_events) ⇒ Object
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 154 def self.analyze_memory_performance(memory_events) return {} if memory_events.empty? allocations = memory_events[:allocations] || [] large_objects = memory_events[:large_objects] || [] snapshots = memory_events[:memory_snapshots] || [] request_allocations = memory_events[:request_allocations] # Calculate memory growth memory_growth = 0 if memory_events[:memory_before] && memory_events[:memory_after] memory_growth = memory_events[:memory_after] - memory_events[:memory_before] end # Calculate allocation totals total_allocations = allocations.sum { |a| a[:count] } total_allocated_size = allocations.sum { |a| a[:size] } if request_allocations total_allocated_size = request_allocations[:allocated_bytes].to_i if total_allocated_size.zero? end gc_allocations = nil if memory_events[:gc_before] && memory_events[:gc_after] gc_allocations = (memory_events[:gc_after][:total_allocated_objects] || 0) - (memory_events[:gc_before][:total_allocated_objects] || 0) end if gc_allocations.to_i > 0 total_allocations = gc_allocations elsif total_allocations.zero? && request_allocations total_allocations = request_allocations[:allocations].to_i end # Group allocations by class allocations_by_class = allocations.group_by { |a| a[:class_name] } .transform_values { |allocs| { count: allocs.sum { |a| a[:count] }, size: allocs.sum { |a| a[:size] } } } # Find top allocating classes top_allocating_classes = allocations_by_class.sort_by { |_, data| -data[:size] }.first(10) # Analyze large objects large_object_analysis = analyze_large_objects(large_objects) # Analyze memory snapshots for trends memory_trends = analyze_memory_trends(snapshots) # Calculate GC efficiency gc_efficiency = calculate_gc_efficiency(memory_events[:gc_before], memory_events[:gc_after]) # Analyze object type deltas (by Ruby object type, not class) object_type_deltas = {} if memory_events[:object_counts_before].is_a?(Hash) && memory_events[:object_counts_after].is_a?(Hash) before = memory_events[:object_counts_before] after = memory_events[:object_counts_after] keys = (before.keys + after.keys).uniq keys.each do |k| object_type_deltas[k] = (after[k] || 0) - (before[k] || 0) end end { memory_growth_mb: memory_growth.round(2), total_allocations: total_allocations, total_allocated_size: total_allocated_size, total_allocated_size_mb: (total_allocated_size / 1_000_000.0).round(2), allocations_per_second: (memory_events[:duration_seconds] > 0) ? (total_allocations.to_f / memory_events[:duration_seconds]).round(2) : 0, top_allocating_classes: top_allocating_classes.map { |class_name, data| { class: class_name, name: class_name, count: data[:count], size: data[:size], size_mb: (data[:size] / 1_000_000.0).round(2) } }, large_objects: large_object_analysis, memory_trends: memory_trends, gc_efficiency: gc_efficiency, memory_snapshots_count: snapshots.count, object_type_deltas: top_object_type_deltas(object_type_deltas, limit: 10) } end |
.analyze_memory_trends(snapshots) ⇒ Object
306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 306 def self.analyze_memory_trends(snapshots) return {} if snapshots.length < 2 # Calculate memory growth rate between snapshots memory_values = snapshots.map { |s| s[:memory_usage] } memory_growth_rates = [] (1...memory_values.length).each do |i| growth = memory_values[i] - memory_values[i - 1] time_diff = snapshots[i][:timestamp] - snapshots[i - 1][:timestamp] rate = (time_diff > 0) ? growth / time_diff : 0 memory_growth_rates << rate end { average_growth_rate_mb_per_second: memory_growth_rates.sum / memory_growth_rates.length, max_growth_rate_mb_per_second: memory_growth_rates.max, memory_volatility: memory_growth_rates.map(&:abs).sum / memory_growth_rates.length, peak_memory_mb: memory_values.max, min_memory_mb: memory_values.min } end |
.calculate_gc_efficiency(gc_before, gc_after) ⇒ Object
329 330 331 332 333 334 335 336 337 338 339 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 329 def self.calculate_gc_efficiency(gc_before, gc_after) return {} unless gc_before && gc_after { gc_count_increase: (gc_after[:count] || 0) - (gc_before[:count] || 0), heap_pages_increase: (gc_after[:heap_allocated_pages] || 0) - (gc_before[:heap_allocated_pages] || 0), objects_allocated: (gc_after[:total_allocated_objects] || 0) - (gc_before[:total_allocated_objects] || 0), gc_frequency: (gc_after[:count] && gc_before[:count]) ? (gc_after[:count] - gc_before[:count]).to_f / [gc_after[:count], 1].max : 0 } end |
.count_objects_snapshot ⇒ Object
296 297 298 299 300 301 302 303 304 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 296 def self.count_objects_snapshot if defined?(ObjectSpace) && ObjectSpace.respond_to?(:count_objects) ObjectSpace.count_objects.dup else {} end rescue {} end |
.current_events ⇒ Object
Current frame (top of stack) for nested job tracking; nil if none.
52 53 54 55 56 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 52 def self.current_events stack = Thread.current[THREAD_LOCAL_KEY] return nil unless stack.is_a?(Array) && stack.any? stack.last end |
.gc_stats ⇒ Object
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 349 def self.gc_stats if defined?(GC) && GC.respond_to?(:stat) stats = GC.stat { count: stats[:count] || 0, heap_allocated_pages: stats[:heap_allocated_pages] || 0, heap_sorted_pages: stats[:heap_sorted_pages] || 0, total_allocated_objects: stats[:total_allocated_objects] || 0, heap_live_slots: stats[:heap_live_slots] || 0, heap_eden_pages: stats[:heap_eden_pages] || 0, heap_tomb_pages: stats[:heap_tomb_pages] || 0 } else {} end rescue {} end |
.heap_pages ⇒ Object
378 379 380 381 382 383 384 385 386 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 378 def self.heap_pages if defined?(GC) && GC.respond_to?(:stat) GC.stat[:heap_allocated_pages] || 0 else 0 end rescue 0 end |
.memory_usage_mb ⇒ Object
341 342 343 344 345 346 347 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 341 def self.memory_usage_mb # MemoryHelpers.rss_mb reads /proc/self/status on Linux and caches for # ~1 second across threads, so this is safe to call per-request. DeadBro::MemoryHelpers.rss_mb rescue 0 end |
.object_count ⇒ Object
368 369 370 371 372 373 374 375 376 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 368 def self.object_count if defined?(GC) && GC.respond_to?(:stat) GC.stat[:heap_live_slots] || 0 else 0 end rescue 0 end |
.object_space_available? ⇒ Boolean
258 259 260 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 258 def self.object_space_available? defined?(ObjectSpace) && ObjectSpace.respond_to?(:each_object) && ObjectSpace.respond_to?(:memsize_of) end |
.record_request_allocations(allocations:, allocated_bytes:) ⇒ Object
Record request-level allocation counters from Rails instrumentation.
97 98 99 100 101 102 103 104 105 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 97 def self.record_request_allocations(allocations:, allocated_bytes:) events = current_events return unless events events[:request_allocations] = { allocations: allocations, allocated_bytes: allocated_bytes } end |
.sample_large_objects ⇒ Object
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 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 262 def self.sample_large_objects results = [] return results unless object_space_available? begin # Sample across common heap object types ObjectSpace.each_object do |obj| # Randomly sample to control overhead next unless rand < LARGE_OBJECT_SAMPLE_RATE size = begin ObjectSpace.memsize_of(obj) rescue 0 end next unless size && size > LARGE_OBJECT_THRESHOLD klass = begin (obj.respond_to?(:class) && obj.class) ? obj.class.name : "Unknown" rescue "Unknown" end results << {class_name: klass, size: size, size_mb: (size / 1_000_000.0).round(2)} break if results.length >= MAX_LARGE_OBJECTS end rescue # Best-effort only end # Sort largest first and keep top N results.sort_by { |h| -h[:size] }.first(MAX_LARGE_OBJECTS) end |
.snapshot_at(label) ⇒ Object
Helper method to take memory snapshots at specific points
389 390 391 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 389 def self.snapshot_at(label) take_memory_snapshot(label) end |
.start_request_tracking ⇒ Object
58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 58 def self.start_request_tracking # Only track if memory tracking is enabled return unless DeadBro.configuration.memory_tracking_enabled frame = { allocations: [], memory_snapshots: [], large_objects: [], request_allocations: nil, gc_before: gc_stats, memory_before: memory_usage_mb, start_time: Time.now.to_f, object_counts_before: count_objects_snapshot } (Thread.current[THREAD_LOCAL_KEY] ||= []) << frame end |
.stop_request_tracking ⇒ Object
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 75 def self.stop_request_tracking stack = Thread.current[THREAD_LOCAL_KEY] events = (stack.is_a?(Array) && stack.any?) ? stack.pop : nil Thread.current[THREAD_LOCAL_KEY] = nil if stack.nil? || stack.empty? if events events[:gc_after] = gc_stats events[:memory_after] = memory_usage_mb events[:end_time] = Time.now.to_f events[:duration_seconds] = events[:end_time] - events[:start_time] events[:object_counts_after] = count_objects_snapshot # Fallback large object detection via ObjectSpace sampling if (events[:large_objects].nil? || events[:large_objects].empty?) && object_space_available? events[:large_objects] = sample_large_objects end end events || {} end |
.subscribe!(client: Client.new) ⇒ Object
Cap number of large objects captured per request
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 22 def self.subscribe!(client: Client.new) # Only enable allocation tracking if explicitly enabled (expensive!) return unless DeadBro.configuration.allocation_tracking_enabled if defined?(ActiveSupport::Notifications) && ActiveSupport::Notifications.notifier.respond_to?(:subscribe) begin # Subscribe to process_action to capture request-level allocation counters ActiveSupport::Notifications.subscribe(PROCESS_ACTION_EVENT) do |*args| event = if args.length == 1 && args.first.is_a?(ActiveSupport::Notifications::Event) args.first else ActiveSupport::Notifications::Event.new(*args) end allocations = event.respond_to?(:allocations) ? event.allocations : event.payload[:allocations] allocated_bytes = event.respond_to?(:allocated_bytes) ? event.allocated_bytes : event.payload[:allocated_bytes] next unless allocations || allocated_bytes record_request_allocations( allocations: allocations, allocated_bytes: allocated_bytes ) end rescue # Allocation tracking might not be available in all Ruby versions end end rescue # Never raise from instrumentation install end |
.take_memory_snapshot(label = nil) ⇒ Object
138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 138 def self.take_memory_snapshot(label = nil) events = current_events return unless events snapshot = { label: label || "snapshot_#{Time.now.to_i}", memory_usage: memory_usage_mb, gc_stats: gc_stats, timestamp: Time.now.utc.to_i, object_count: object_count, heap_pages: heap_pages } events[:memory_snapshots] << snapshot end |
.top_object_type_deltas(deltas, limit: 10) ⇒ Object
253 254 255 256 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 253 def self.top_object_type_deltas(deltas, limit: 10) return {} unless deltas.is_a?(Hash) deltas.sort_by { |_, v| -v.abs }.first(limit).to_h end |
.track_allocation(data, started, finished) ⇒ Object
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 |
# File 'lib/dead_bro/memory_tracking_subscriber.rb', line 107 def self.track_allocation(data, started, finished) events = current_events return unless events # Only track if we have meaningful allocation data return unless data.is_a?(Hash) && data[:count] && data[:size] # Limit allocations per request to prevent memory bloat allocations = events[:allocations] return if allocations.length >= MAX_ALLOCATIONS_PER_REQUEST # Simplified allocation tracking (avoid expensive operations) allocation = { class_name: data[:class_name] || "Unknown", count: data[:count], size: data[:size] # Removed expensive fields: duration_ms, timestamp, memory_usage } # Track large object allocations (these are rare and important) if data[:size] > LARGE_OBJECT_THRESHOLD large_object = allocation.merge( large_object: true, size_mb: (data[:size] / 1_000_000.0).round(2) ) events[:large_objects] << large_object end events[:allocations] << allocation end |