Running GoodJob in production with systemd throws an error due to a wrong communication with WatchDog

[tagrudev]

Hello, I am trying to run good_job in production using the systemd service example that was added in #1029, unfortunately there's a problem with the communication between good_job and watchdog.

My setup

ruby 3.2.2
good_job 3.21.2 
rails 7.1.1
ubuntu 22.04

My goodjob service running as a user service

Description=GoodJob Background Job Processor
After=postgresql.service network.target

[Service]
Type=notify
WatchdogSec=5s

WorkingDirectory=/my/working/directory

Environment=RAILS_ENV=production
ExecStart=/bin/bash -lc 'exec /usr/local/rbenv/shims/bundle exec good_job start'

RestartSec=1s
Restart=always

StandardOutput=journal
StandardError=journal
SyslogIdentifier=goodjob

[Install]
WantedBy=default.target

This results in

Goodjob is running, but I think there's a problem with the communication with Watchdog.
I think the problem comes from https://github.com/bensheldon/good_job/pull/1029/files#diff-121a5f6ab045b7079f271a0b111a46097545d1d13550bc5b1394d3e7adb94bfbR178

Looking at the other methods it's missing return unless exception is that the problem? adding this locally fixes the issue, but I am not sure if the watchdog communication is ok.

[tagrudev]

I guess

good_job/lib/good_job/systemd_service.rb

Lines 16 to 17 in d06ed8e

	ActiveSupport::Notifications.instrument("systemd_watchdog_error.good_job", { error: thread_error })
	GoodJob._on_thread_error(thread_error) if thread_error

This part should be

if thread_error
  ActiveSupport::Notifications.instrument("systemd_watchdog_error.good_job", { error: thread_error })
  GoodJob._on_thread_error(thread_error)
end

[bensheldon]

@tagrudev Nice find! Can you make a PR for that?

[tagrudev]

Hey @bensheldon - yes, already did #1173

[tagrudev]

I am closing this issue since the PR is now merged.