Sha256: 233bc018374e92f2612fafb6794cfd04f251517c38082793e96264e0c4efd0b9
Contents?: true
Size: 1.79 KB
Versions: 14
Compression:
Stored size: 1.79 KB
Contents
namespace :delayed do
task delayed_environment: :environment do
Delayed::Worker.min_priority = ENV['MIN_PRIORITY'].to_i if ENV.key?('MIN_PRIORITY')
Delayed::Worker.max_priority = ENV['MAX_PRIORITY'].to_i if ENV.key?('MAX_PRIORITY')
Delayed::Worker.queues = [ENV['QUEUE']] if ENV.key?('QUEUE')
Delayed::Worker.queues = ENV['QUEUES'].split(',') if ENV.key?('QUEUES')
Delayed::Worker.sleep_delay = ENV['SLEEP_DELAY'].to_i if ENV.key?('SLEEP_DELAY')
Delayed::Worker.read_ahead = ENV['READ_AHEAD'].to_i if ENV.key?('READ_AHEAD')
Delayed::Worker.max_claims = ENV['MAX_CLAIMS'].to_i if ENV.key?('MAX_CLAIMS')
next unless defined?(Rails.application.config)
# By default, Rails < 6.1 overrides eager_load to 'false' inside of rake tasks, which is not ideal in production environments.
# Additionally, the classic Rails autoloader is not threadsafe, so we do not want any autoloading after we start the worker.
# While the zeitwork autoloader technically does not need this workaround, we will still eager load for consistency's sake.
# We will use the cache_classes config as a proxy for determining if we should eager load before booting workers.
if !Rails.application.config.respond_to?(:rake_eager_load) && Rails.application.config.cache_classes
Rails.application.config.eager_load = true
Rails::Application::Finisher.initializers
.find { |i| i.name == :eager_load! }
.bind(Rails.application)
.run
end
end
desc 'start a delayed worker'
task work: :delayed_environment do
Delayed::Worker.new.start
end
desc 'monitor job queue and emit metrics at an interval'
task monitor: :delayed_environment do
Delayed::Monitor.new.start
end
end
# For backwards compatibility:
namespace :jobs do
task work: %i(delayed:work)
end
Version data entries
14 entries across 14 versions & 1 rubygems