perf_results_0.1.0.md

Performance results Jiggler 0.1.0

The tests were run on local (Ubuntu 22.04, Intel(R) Core(TM) i7 6700HQ 2.60GHz). \

Ruby 3.2.0
Redis 7.0.7
Poller interval 5s
Monitoring interval 10s
Logging level WARN

Noop task measures

def perform
  # just an empty job doing nothing
end

AMO - at most once
ALO - at least once

Job Processor	Concurrency	Jobs	Time	Start RSS	Finish RSS
Jiggler 0.1.0 (AMO)	10	100_000	14.17 sec	82_152 kb	100_160 kb
Jiggler 0.1.0 (ALO)	10	100_000	26.56 sec	82_212 kb	103_420 kb
-
Jiggler 0.1.0 (AMO)	50	100_000	13.96 sec	82_060 kb	101_480 kb
Jiggler 0.1.0 (ALO)	50	100_000	26.19 sec	82_204 kb	75_068 kb (GC)

Note: noop task isn't really representative though, it's best to measure performance on some real tasks, to see how much context switching affects specific payload depending on CPU/IO balance.

NET/HTTP requests

Spin-up a local sinatra app to exclude network issues while testing HTTP requests (it uses falcon web server).

require "sinatra"

class MyApp < Sinatra::Base
  get "/hello" do
    sleep(0.2)
    "Hello World!"
  end
end

Then, the code which is going to be performed within the workers should make a net/http request to the local endpoint.

# a single job takes ~0.21s to perform
def perform
  uri = URI("http://127.0.0.1:9292/hello")
  res = Net::HTTP.get_response(uri)
  puts "Request Error!!!" unless res.is_a?(Net::HTTPSuccess)
end

Job Processor	Concurrency	Jobs	Time	Start RSS	Finish RSS	%CPU
Jiggler 0.1.0 (AMO)	10	1_000	23.25 sec	28_872 kb	44_896 kb	10.8%
Jiggler 0.1.0 (ALO)	10	1_000	22.86 sec	28_868 kb	45_524 kb	11.3%
-
Jiggler 0.1.0 (AMO)	50	1_000	6.78 sec	28_804 kb	46_740 kb	36.9%
Jiggler 0.1.0 (ALO)	50	1_000	6.45 sec	28_948 kb	47_624 kb	38.1%
-
Jiggler 0.1.0 (AMO)	100	1_000	4.71 sec	29_464 kb	49_040 kb	55.7%
Jiggler 0.1.0 (ALO)	100	1_000	4.49 sec	28_948 kb	47_624 kb	60.0%

PostgreSQL connection/queries

pg gem supports Ruby's Fiber.scheduler starting from 1.3.0 version. Make sure yours DB-adapter supports it.

### global namespace
require "pg"

$pg_pool = ConnectionPool.new(size: CONCURRENCY) do
  PG.connect({ dbname: "test", password: "test", user: "test" })
end

### worker context
# a single job takes ~0.102s to perform
def perform
  $pg_pool.with do |conn|
    conn.exec("SELECT pg_sleep(0.1)")
  end
end

Job Processor	Concurrency	Jobs	Time	Start RSS	Finish RSS	%CPU
Jiggler 0.1.0 (AMO)	10	1_000	12.95 sec	28_872 kb	50_016 kb	12.4%
Jiggler 0.1.0 (ALO)	10	1_000	12.99 sec	28_868 kb	50_076 kb	12.6%
-
Jiggler 0.1.0 (AMO)	50	1_000	5.38 sec	28_724 kb	51_904 kb	47.0%
Jiggler 0.1.0 (ALO)	50	1_000	5.37 sec	28_908 kb	52_516 kb	50.7%
-
Jiggler 0.1.0 (AMO)	100	1_000	5.39 sec	29_464 kb	54_644 kb	69.8%
Jiggler 0.1.0 (ALO)	100	1_000	5.17 sec	28_896 kb	56_148 kb	70.3%

File IO

def perform(file_name, id)
  File.open(file_name, "a") { |f| f.write("#{id}\n") }
end

Job Processor	Concurrency	Jobs	Time	Start RSS	Finish RSS	%CPU
Jiggler 0.1.0 (AMO)	10	50_000	11.95 sec	54_760 kb	56_304 kb	91.0%
Jiggler 0.1.0 (ALO)	10	50_000	18.07 sec	54_780 kb	58_128 kb	91.6%
-
Jiggler 0.1.0 (AMO)	50	50_000	10.9 sec	54_608 kb	57_432 kb	91.5%
Jiggler 0.1.0 (ALO)	50	50_000	17.18 sec	54_748 kb	60_128 kb	96.8%

CPU-only job

def fib(n)
  if n <= 1
    1
  else
    (fib(n-1) + fib(n-2))
  end
end

# a single job takes ~0.035s to perform
def perform(_idx)
  fib(25)
end

Job Processor	Concurrency	Jobs	Time	Start RSS	Finish RSS	%CPU
Jiggler 0.1.0 (AMO)	10	1_000	35.48 sec	26_916 kb	43_460 kb	98.1%
Jiggler 0.1.0 (ALO)	10	1_000	35.45 sec	27_100 kb	43_860 kb	94.5%
-
Jiggler 0.1.0 (AMO)	50	1_000	35.77 sec	27_188 kb	44_532 kb	98.3%
Jiggler 0.1.0 (ALO)	50	1_000	35.93 sec	27_020 kb	45_816 kb	99.7%

Idle

3 minutes of idle work.

Job Processor	Concurrency	Start RSS	Finish RSS	%CPU
Jiggler 0.1.0 (AMO)	10	27_508 kb	44_568 kb	1.0%
Jiggler 0.1.0 (ALO)	10	27_404 kb	43_224 kb	1.0%
-
Jiggler 0.1.0 (AMO)	50	27_184 kb	44_952 kb	1.3%
Jiggler 0.1.0 (ALO)	50	27_180 kb	44_924 kb	1.0%
-
Jiggler 0.1.0 (AMO)	100	27_508 kb	46_352 kb	1.6%
Jiggler 0.1.0 (ALO)	100	26_640 kb	47_172 kb	1.0%