10

I'm running Puma 2.8.2 server for stubbing some of my back end services. Occasionally Puma server just dies without any reason.

There is no errors in error.log And here is excerpt from access.log:

10.210.140.21 - - [15/Oct/2014 09:28:22] "GET /status HTTP/1.1" 200 - 0.0030
- Gracefully stopping, waiting for requests to finish
- Goodbye!
=== puma startup: 2014-10-15 09:28:24 +0100 ===
10.210.140.21 - - [15/Oct/2014 09:28:24] "GET /status HTTP/1.1" 200 - 0.0254

Any thoughts what could cause this?

LukasMac
  • 848
  • 1
  • 8
  • 20
  • 1
    Maybe you question has already been answered: http://stackoverflow.com/questions/23867784/puma-rails-on-linux-restart-when-process-dies – SlySherZ Oct 27 '14 at 16:50
  • 1
    Thanks @SlySherZ. I saw this post, unfortunately it doesn't tell how to prevent puma from dying. Just how to start/stop/restart it. – LukasMac Oct 27 '14 at 17:04
  • You could try patching Puma to output more useful logs: https://github.com/puma/puma/issues/1273 – brainbag Nov 17 '17 at 13:24

1 Answers1

1

If this problem is becoming in Heroku and you are using Heroku free dyno, the problem is that you use the free dyno. Because your app goes to sleep after 30 minutes of inactivity.

You can find more info at the following link:

https://devcenter.heroku.com/articles/free-dyno-hours

Hamdi Bayhan
  • 1,773
  • 4
  • 17
  • 20