Auto DevOps can't start unicorn process

Deployed pods crash on startup because the unicorn process thinks it is already running.

Using this config file: https://gitlab.com/devin/version-gitlab-com/blob/master/config/unicorn/production.rb

/app/vendor/bundle/ruby/2.5.0/gems/unicorn-5.5.1/lib/unicorn/http_server.rb:207:in `pid=': Already running on PID:13 (or pid=tmp/pids/server.pid is stale) (ArgumentError)
	from /app/vendor/bundle/ruby/2.5.0/gems/unicorn-5.5.1/lib/unicorn/http_server.rb:139:in `start'
	from /app/vendor/bundle/ruby/2.5.0/gems/unicorn-rails-2.2.1/lib/unicorn_rails.rb:33:in `run'
	from /app/vendor/bundle/ruby/2.5.0/gems/rack-2.0.7/lib/rack/server.rb:297:in `start'
	from /app/vendor/bundle/ruby/2.5.0/gems/railties-5.0.7.1/lib/rails/commands/server.rb:104:in `start'
	from /app/vendor/bundle/ruby/2.5.0/gems/railties-5.0.7.1/lib/rails/commands/commands_tasks.rb:90:in `block in server'
	from /app/vendor/bundle/ruby/2.5.0/gems/railties-5.0.7.1/lib/rails/commands/commands_tasks.rb:85:in `tap'
	from /app/vendor/bundle/ruby/2.5.0/gems/railties-5.0.7.1/lib/rails/commands/commands_tasks.rb:85:in `server'
	from /app/vendor/bundle/ruby/2.5.0/gems/railties-5.0.7.1/lib/rails/commands/commands_tasks.rb:49:in `run_command!'
	from /app/vendor/bundle/ruby/2.5.0/gems/railties-5.0.7.1/lib/rails/commands.rb:18:in `<top (required)>'
	from bin/rails:4:in `require'
	from bin/rails:4:in `<main>'

It would be nice to have some documentation on the preferred way to expose a web service on a pod running a rails app