Module: LinuxStat::ProcessInfo

Defined in:
lib/linux_stat/process_info.rb

Overview

Shows various information about a process that is either running, sleeping, idle or a zombie. Most methods can take a PID, but some uses polling to calculate something, they can accept options instead of arguments. Consult the documentation on the specific methods for more details on that.

Class Method Summary collapse

Class Method Details

.cmdline(pid = $$) ⇒ Object

cmdline(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the total command of the process.

The output is String. For example:

LinuxStat::ProcessInfo.cmdline

"ruby -r linux_stat -e p LinuxStat::ProcessInfo.cmdline"

If the info isn’t available it will return an empty frozen String.



61
62
63
64
65
66
67
68
 
# File 'lib/linux_stat/process_info.rb', line 61

def cmdline(pid = $$)
  file = "/proc/#{pid}/cmdline".freeze
  return ''.freeze unless File.readable?(file)

  _cmdline = IO.read(file)
  _cmdline.gsub!(?\u0000, ?\s)
  _cmdline.tap(&:strip!)
end

.command_name(pid = $$) ⇒ Object

command_name(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the total command name of the process.

The output is String. For example:

LinuxStat::ProcessInfo.command_name

"ruby"

If the info isn’t available it will return an empty frozen String.



85
86
87
88
89
90
91
92
93
 
# File 'lib/linux_stat/process_info.rb', line 85

def command_name(pid = $$)
  # Do note that the /proc/ppid/comm may not contain the full name
  file = "/proc/#{pid}/cmdline".freeze
  return ''.freeze unless File.readable?(file)

  _cmdline = IO.read(file)
  _cmdline.gsub!(?\u0000, ?\s)
  File.split(_cmdline.tap(&:strip!).split[0])[-1]
end

.cpu_stat(pid: $$, sleep: ticks_to_ms_t5) ⇒ Object

cpu_stat(pid: $$, sleep: 1.0 / LinuxStat::Sysconf.sc_clk_tck * 5)

Where pid is the process ID and sleep time is the interval between measurements.

By default it is the id of the current process ($$), and sleep is LinuxStat::Sysconf.sc_clk_tck

The smallest amount of available sleep time is 1.0 / LinuxStat::Sysconf.sc_clk_tck * 5.

  • Note 1:

  1. Do note that the sleep time can slow down your application.

  2. And it’s only needed for the cpu usage calculation.

It retuns the CPU usage, threads, and the last executed CPU in Hash.

For example:

LinuxStat::ProcessInfo.cpu_stat

=> {:cpu_usage=>0.0, :threads=>1, :last_executed_cpu=>1}

But if the info isn’t available, it will return an empty Hash.

The :cpu_usage is in percentage. It’s also divided with the number of CPU.

:cpu_usage for example, will return 25.0 if the CPU count is 4, and the process is using 100% of a thread / core.

A value of 100.0 indicates it is using 100% processing power available to the system.

The :threads returns the number of threads for the process. The value is a Integer.

  • Note 2:

  1. If you just need the CPU usage run LinuxStat::ProcessInfo.cpu_usage(pid = $$)

  2. If you just need the threads run LinuxStat::ProcessInfo.threads(pid = $$)

  3. If you just need the last executed CPU run LinuxStat::ProcessInfo.last_executed_cpu(pid = $$)

  4. Running this method is slower and it opens multiple files at once

Only use this method if you need all of the data at once, in such case, it’s more efficient to use this method.

The :last_executed_cpu also returns an Integer indicating the last executed cpu of the process.



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
 
# File 'lib/linux_stat/process_info.rb', line 254

def cpu_stat(pid: $$, sleep: ticks_to_ms_t5)
  file = "/proc/#{pid}/stat"
  ticks = get_ticks

  return {} unless File.readable?(file)
  utime, stime, starttime = IO.read(file)
    .split.values_at(13, 14, 21).map(&:to_f)

  uptime = IO.read('/proc/uptime'.freeze).to_f * ticks

  total_time = utime + stime
  idle1 = uptime - starttime - total_time

  sleep(sleep)

  return {} unless File.readable?(file)
  stat = IO.read(file).split

  utime2, stime2, starttime2 = stat.values_at(13, 14, 21).map(&:to_f)
  uptime = IO.read('/proc/uptime'.freeze).to_f * ticks

  total_time2 = utime2 + stime2
  idle2 = uptime - starttime2 - total_time2

  totald = idle2.+(total_time2).-(idle1 + total_time)
  cpu_u = totald.-(idle2 - idle1).fdiv(totald).abs.*(100)./(cpu_count)

  {
    cpu_usage: cpu_u > 100 ? 100.0 : cpu_u.round(2),
    threads: stat[19].to_i,
    last_executed_cpu: stat[38].to_i
  }
end

.cpu_usage(pid: $$, sleep: ticks_to_ms_t5) ⇒ Object

cpu_usage(pid: $$, sleep: 1.0 / LinuxStat::Sysconf.sc_clk_tck * 5)

Where pid is the process ID and sleep time is the interval between measurements.

By default it is the id of the current process ($$), and sleep is 1.0 / LinuxStat::Sysconf.sc_clk_tck * 5

The smallest amount of available sleep time is LinuxStat::Sysconf.sc_clk_tck.

It retuns the CPU usage as Float.

For example:

LinuxStat::ProcessInfo.cpu_usage

=> 10.0

10.0 means it’s using 10% of the total processing power of the system.

The value is divided with the configured number of CPU and not online CPU.

A value of 100.0 indicates it is using 100% processing power available to the system.

But if the info isn’t available, it will return nil.

This method is more efficient than running LinuxStat::ProcessInfo.cpu_stat()



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
 
# File 'lib/linux_stat/process_info.rb', line 314

def cpu_usage(pid: $$, sleep: ticks_to_ms_t5)
  file = "/proc/#{pid}/stat"
  ticks = get_ticks

  return nil unless File.readable?(file)
  utime, stime, starttime = IO.read(file)
    .split.values_at(13, 14, 21).map(&:to_f)

  uptime = IO.read('/proc/uptime'.freeze).to_f * ticks

  total_time = utime + stime
  idle1 = uptime - starttime - total_time

  sleep(sleep)

  return nil unless File.readable?(file)
  utime2, stime2, starttime2 = IO.read(file)
    .split.values_at(13, 14, 21).map(&:to_f)

  uptime = IO.read('/proc/uptime'.freeze).to_f * ticks

  total_time2 = utime2 + stime2
  idle2 = uptime - starttime2 - total_time2

  totald = idle2.+(total_time2).-(idle1 + total_time)

  u = totald.-(idle2 - idle1).fdiv(totald).abs.*(100)./(cpu_count)
  u > 100 ? 100.0 : u.round(2)
end

.gid(pid = $$) ⇒ Object

gid(pid = $$)

returns the GIDs of the process as an Hash containing the following data:

:real, :effective, :saved_set, :filesystem_uid

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return an empty Hash.



477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
 
# File 'lib/linux_stat/process_info.rb', line 477

def gid(pid = $$)
  file = "/proc/#{pid}/status".freeze
  return nil unless File.readable?(file)

  data = IO.foreach(file.freeze).find { |x|
    x[/Gid.*\d*/]
  }.split.drop(1)

  {
    real: data[0].to_i,
    effective: data[1].to_i,
    saved_set: data[2].to_i,
    filesystem_uid: data[3].to_i
  }
end

.last_executed_cpu(pid = $$) ⇒ Object

last_executed_cpu(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the last executed CPU in Integer.

For example:

LinuxStat::ProcessInfo.last_executed_cpu

=> 2

But if the info isn’t available, it will return nil.

This method is way more efficient than running LinuxStat::ProcessInfo.cpu_stat()



440
441
442
443
444
445
 
# File 'lib/linux_stat/process_info.rb', line 440

def last_executed_cpu(pid = $$)
  file = "/proc/#{pid}/stat".freeze
  return nil unless File.readable?(file)

  IO.read(file).split[38].to_i
end

.mem_stat(pid = $$) ⇒ Object

mem_stat(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the memory, virtual memory, and resident memory of the process.

All values are in kilobytes.

The output is a Hash. For example:

LinuxStat::ProcessInfo.mem_stat

{:memory=>8515.584, :virtual_memory=>79781.888, :resident_memory=>13955.072}

  • Note:

  1. If you need only memory usage of a process, run LinuxStat::ProcessInfo.memory(pid)

  2. If you need only virtual memory for a process, run LinuxStat::ProcessInfo.virtual_memory(pid)

  3. If you need only resident memory of a process, run LinuxStat::ProcessInfo.resident_memory(pid)

This method opens opens multiple files.

But if you need all of the info, then running this method once is efficient.

If the info isn’t available it will return an empty Hash.



121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
 
# File 'lib/linux_stat/process_info.rb', line 121

def mem_stat(pid = $$)
  statm = "/proc/#{pid}/statm".freeze
  return {} unless File.readable?(statm)

  data = IO.read(statm).split

  _rss_anon = (data[1] && data[2]) ? data[1].to_i.-(data[2].to_i).*(pagesize).fdiv(1000) : nil
  _virtual_memory = data[0] ? data[0].to_i*(pagesize).fdiv(1000) : nil
  _resident_memory = data[1] ? data[1].to_i.*(pagesize).fdiv(1000) : nil

  {
    memory: _rss_anon,
    virtual_memory: _virtual_memory,
    resident_memory: _resident_memory
  }
end

.memory(pid = $$) ⇒ Object

memory(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the memory of the process. The value is in kilobytes.

The output is an Integer. For example:

LinuxStat::ProcessInfo.memory

8523.776

If the info isn’t available it will return nil.



154
155
156
157
158
159
160
 
# File 'lib/linux_stat/process_info.rb', line 154

def memory(pid = $$)
  file = "/proc/#{pid}/statm".freeze
  return nil unless File.readable?(file)

  data = IO.read(file).split
  (data[1] && data[2]) ? data[1].to_i.-(data[2].to_i).*(pagesize).fdiv(1000) : nil
end

.nice(pid = $$) ⇒ Object

nice(pid = $$)

Returns the nice of the process

The output value is an Integer ranging from -20 to 19

-20 means the process has high priority, and 19 means the process has low priority

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return nil.



614
615
616
617
618
619
 
# File 'lib/linux_stat/process_info.rb', line 614

def nice(pid = $$)
  file = "/proc/#{pid}/stat"
  return nil unless File.readable?(file)

  IO.foreach(file, ' ').first(19)[-1].to_i
end

.nproc(pid = $$) ⇒ Object Also known as: count_cpu

nproc(pid = $$)

Returns the cpu allocated to the process.

The output value is an Integer.

For example:

$ taskset -c 0 irb

irb(main):001:0> require 'linux_stat'

=> true

irb(main):002:0> LinuxStat::ProcessInfo.nproc

=> 1

irb(main):003:0> LinuxStat::ProcessInfo.nproc 11562

=> 3

irb(main):004:0> LinuxStat::ProcessInfo.nproc 12513

=> 4

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return nil.



647
648
649
 
# File 'lib/linux_stat/process_info.rb', line 647

def nproc(pid = $$)
  LinuxStat::Nproc.count_cpu_for_pid(pid)
end

.owner(pid = $$) ⇒ Object

owner(pid = $$)

Returns the owner of the process But if the status is not available, it will return an empty frozen String.



498
499
500
501
502
503
504
505
506
507
 
# File 'lib/linux_stat/process_info.rb', line 498

def owner(pid = $$)
  file = "/proc/#{pid}/status".freeze
  return ''.freeze unless File.readable?(file)

  gid = IO.foreach(file.freeze).find { |x|
    x[/Gid.*\d*/]
  }.split.drop(1)[2].to_i

  LinuxStat::User.username_by_gid(gid)
end

.resident_memory(pid = $$) ⇒ Object

resident_memory(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the resident memory for the process.

The value is in kilobytes.

The output is an Integer. For example:

LinuxStat::ProcessInfo.cpu_stat

=> 13996.032

If the info isn’t available it will return nil.



204
205
206
207
208
209
210
 
# File 'lib/linux_stat/process_info.rb', line 204

def resident_memory(pid = $$)
  file = "/proc/#{pid}/statm".freeze
  return nil unless File.readable?(file)

  _vm_rss = IO.read(file).split[1]
  _vm_rss ? _vm_rss.to_i.*(pagesize).fdiv(1000) : nil
end

.running_time(pid = $$) ⇒ Object

running_time(pid = $$)

Returns the time (in seconds, as Float) the process is running for.

For example:

LinuxStat::ProcessInfo.running_time 14183

=> 1947.61

It always rounds the float number upto 2 decimal places

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return nil.



570
571
572
573
574
575
576
577
578
579
 
# File 'lib/linux_stat/process_info.rb', line 570

def running_time(pid = $$)
  stat_file = "/proc/#{pid}/stat".freeze
  uptime = "/proc/uptime".freeze

  @@u_readable ||= File.readable?(uptime)
  return nil unless @@u_readable && File.readable?(stat_file)

  IO.foreach(uptime, ' '.freeze).next.to_f
    .-(IO.read(stat_file).split[21].to_f / get_ticks).round(2)
end

.start_time(pid = $$) ⇒ Object

start_time(pid = $$)

Returns the time (as Time object) the process was started.

For example:

LinuxStat::ProcessInfo.start_time 14183

=> 2020-12-16 13:31:43 +0000

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return nil.

The timezone returned based on current TZ. Thus the timezone could be affected by changing the ENV variable.

Don’t trust the timezone returned by the time.



551
552
553
554
555
 
# File 'lib/linux_stat/process_info.rb', line 551

def start_time(pid = $$)
  # Getting two Time objects and dealing with floating point numbers
  # Just to make sure the time goes monotonically
  Time.at(start_time_epoch(pid))
end

.start_time_epoch(pid = $$) ⇒ Object

start_time_epoch(pid = $$)

Returns the epoch time (as Integer) the process was started.

For example:

LinuxStat::ProcessInfo.start_time_epoch 526

=> 1608097744

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return nil.



520
521
522
523
524
525
526
527
528
529
530
531
532
533
 
# File 'lib/linux_stat/process_info.rb', line 520

def start_time_epoch(pid = $$)
  stat_file = "/proc/#{pid}/stat".freeze
  uptime = "/proc/uptime".freeze

  @@u_readable ||= File.readable?(uptime)
  return nil unless @@u_readable && File.readable?(stat_file)

  u = IO.foreach(uptime, ' '.freeze).next.to_f
  st = (IO.foreach(stat_file, ' '.freeze).first(22)[-1].to_f / get_ticks)

  # Getting two Time objects and dealing with floating point numbers
  # Just to make sure the time goes monotonically
  Time.now.-(u - st).to_i
end

.state(pid = $$) ⇒ Object

state(pid = $$)

Returns the state of the process as a frozen String

  • A process could have multiple states:

  1. S => Sleeping

  2. R => Running

  3. I => Idle

  4. Z => Zombie

It returns any one of them.

If the info isn’t available or the argument passed doesn’t exist as a process ID, it will return an empty String.



599
600
601
602
603
 
# File 'lib/linux_stat/process_info.rb', line 599

def state(pid = $$)
  file = "/proc/#{pid}/stat".freeze
  return ''.freeze unless File.readable?(file)
  IO.foreach(file, ' '.freeze).first(3)[-1].tap(&:rstrip!).freeze
end

.thread_usage(pid: $$, sleep: ticks_to_ms_t5) ⇒ Object

thread_usage(pid: $$, sleep: 1.0 / LinuxStat::Sysconf.sc_clk_tck * 5)

Where pid is the process ID and sleep time is the interval between measurements.

By default it is the id of the current process ($$), and sleep is 1.0 / LinuxStat::Sysconf.sc_clk_tck * 5

The smallest amount of available sleep time is LinuxStat::Sysconf.sc_clk_tck.

It retuns the per core CPU usage as Float.

For example:

LinuxStat::ProcessInfo.core_usage

=> 200.0

A value of 100.0 indicates it is using 100% processing power of a core.

The value could be 0 to (100 * the number of CPU threads (including hyperthreading) in the system)

But if the info isn’t available, it will return nil.



366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
 
# File 'lib/linux_stat/process_info.rb', line 366

def thread_usage(pid: $$, sleep: ticks_to_ms_t5)
  file = "/proc/#{pid}/stat"
  ticks = get_ticks

  return nil unless File.readable?(file)
  utime, stime, starttime = IO.read(file)
    .split.values_at(13, 14, 21).map(&:to_f)

  uptime = IO.read('/proc/uptime'.freeze).to_f * ticks

  total_time = utime + stime
  idle1 = uptime - starttime - total_time

  sleep(sleep)

  return nil unless File.readable?(file)
  utime2, stime2, starttime2 = IO.read(file)
    .split.values_at(13, 14, 21).map(&:to_f)

  uptime = IO.read('/proc/uptime'.freeze).to_f * ticks

  total_time2 = utime2 + stime2
  idle2 = uptime - starttime2 - total_time2

  totald = idle2.+(total_time2).-(idle1 + total_time)

  u = totald.-(idle2 - idle1).fdiv(totald).abs.*(100)

  cpu_count_t100 = cpu_count * 100
  u > cpu_count_t100 ? cpu_count_t100 : u.round(2)
end

.threads(pid = $$) ⇒ Object

threads(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the threads for the current process in Integer.

For example:

LinuxStat::ProcessInfo.threads

=> 2

But if the info isn’t available, it will return nil.

This method is way more efficient than running LinuxStat::ProcessInfo.cpu_stat()



415
416
417
418
419
420
421
 
# File 'lib/linux_stat/process_info.rb', line 415

def threads(pid = $$)
  file = "/proc/#{pid}/stat".freeze
  return nil unless File.readable?(file)

  data = IO.foreach(file, ' '.freeze).first(20)[-1]
  data ? data.to_i : nil
end

.total_io(pid = $$) ⇒ Object

total_io(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the total read/write caused by a process.

The output is Hash.

For example:

LinuxStat::ProcessInfo.total_io

{:read_bytes=>0, :write_bytes=>0}

The output is only based on the total disk IO the process has done.

If the info isn’t available it will return an empty Hash.



30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
 
# File 'lib/linux_stat/process_info.rb', line 30

def total_io(pid = $$)
  return {} unless File.readable?("/proc/#{pid}/io".freeze)
  out = {}

  IO.readlines("/proc/#{pid}/io".freeze).each { |x|
    x.strip!

    if x[/^(read|write)_bytes:\s*\d*$/]
      splitted = x.split
      out.merge!(splitted[0].split(?:)[0].to_sym => splitted[-1].to_i)
    end
  }

  out
end

.uid(pid = $$) ⇒ Object

uid(pid = $$)

returns the UIDs of the process as an Array of Integers.

If the info isn’t available it returns an empty Array.



453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
 
# File 'lib/linux_stat/process_info.rb', line 453

def uid(pid = $$)
  file = "/proc/#{pid}/status".freeze
  return nil unless File.readable?(file)

  data = IO.foreach(file.freeze).find { |x|
    x[/Uid.*\d*/]
  }.to_s.split.drop(1)

  {
    real: data[0].to_i,
    effective: data[1].to_i,
    saved_set: data[2].to_i,
    filesystem_uid: data[3].to_i
  }
end

.virtual_memory(pid = $$) ⇒ Object

virtual_memory(pid = $$)

Where pid is the process ID.

By default it is the id of the current process ($$)

It retuns the virtual memory for the process.

The value is in kilobytes.

The output is an Integer. For example:

LinuxStat::ProcessInfo.virtual_memory

79781.888

If the info isn’t available it will return nil.



179
180
181
182
183
184
185
 
# File 'lib/linux_stat/process_info.rb', line 179

def virtual_memory(pid = $$)
  file = "/proc/#{pid}/statm".freeze
  return nil unless File.readable?(file)

  _virtual_memory = IO.read(file).split[0]
  _virtual_memory ? _virtual_memory.to_i.*(pagesize).fdiv(1000) : nil
end