High memory consumption on HTTParty, question asked on StackOverflow: https://stackoverflow.com/questions/53169751/ruby-memory-usage-gone-wild

memory-test-httparty.rb

require 'bundler/inline'
require 'json'

gemfile do
  source 'https://rubygems.org'
  gem 'httparty'
  gem 'nokogiri'
  gem 'memory_profiler'
end

xml_base_url = 'http://noneed.dk/test.xml?date='
auth_end_point = 'http://noneed.dk/session.php'
auth_payload = {"user" => "hi@noneed.dk", "password" => "secret"}
image_search_base_url = 'http://noneed.dk/images.php?id='
image_base_url = 'http://noneed.dk/'


logger = Logger.new(STDOUT)
download_size = 0
time_start = Time.now
requests = 0

report = MemoryProfiler.report do
  # Authenticates
  response = HTTParty.post(auth_end_point, body: {payload: auth_payload})
  cookie_jar = HTTParty::CookieHash.new
  response.headers.get_fields('Set-Cookie').each { |c| cookie_jar.add_cookies(c) }
  logger.debug "Got a cookiejar - #{cookie_jar.to_cookie_string}"
  download_size += response.body.bytesize
  requests += 1

  # Run X days of data
  1.times do |n|
    date = (Time.now + n * 86400).strftime("%F")
    logger.info "Downloads data for #{date}"
    response = HTTParty.get(xml_base_url + date)
    download_size += response.body.bytesize
    requests += 1
    logger.debug "Download completed with code #{response.code} with a total size of #{response.body.bytesize/1024}Kb."

    # Parse downloaded XML and process data items
    xml = Nokogiri::XML(response.body)
    items = xml.xpath("//noneed/item")
    logger.debug "-- found #{items.count} records in the file"

    items.each do |item|
      id = item.at("id").content
      unless id.nil?
        logger.info "Find images for id:#{id}"
        images = HTTParty.post(image_search_base_url + id, cookies: cookie_jar)
        download_size += images.body.bytesize
        requests += 1
        logger.debug "Download completed with code #{images.code} with a total size of #{images.body.bytesize}."
        image_json = JSON.parse(images.body)
        image_json['ids'].each do |img_id|
          logger.debug "-- downloading image #{img_id}"

          image = HTTParty.post(image_base_url + img_id.to_s, cookies: cookie_jar)
          download_size += image.body.bytesize
          requests += 1
          logger.debug "---- download status: #{image.code} with a total size of #{image.body.bytesize/1024.0/1024.0}Mb."
        end
      end
    end
  end

  logger.info "Total download size: #{(download_size / 1024.0 / 1024.0).round(2)}Mb in #{Time.now - time_start} seconds through #{requests} requests."
end

report.pretty_print

logger.info "Completed!"

Identical to https://gist.github.com/mtrolle/96f55822122ecabd3cc46190a6dc18a5 but this version uses HTTParty over RestClient as it was suggested on https://stackoverflow.com/questions/53169751/ruby-memory-usage-gone-wild that HTTParty was a more "commonly used" gem.