If you were using the api, and some external monitoring, you should be able to make something to do this; but simply restarting your nodes over and over won't make a problem go away.

This will just wash away any evidence of what is happening, and result in things never being fixed.

I would highly not recommend doing this as restarting a system does not fix issues. If your system randomly hangs, fix the issue versus having something restarts as restarting the system will put unnecessary load on DigitalOcean if the network to you droplet drops, etc.