My STEEM related bots are actually long-running processes. They sometimes stop, fail and go offline. It might be an unhandled exception, a network issue, a temporary steem public node issue, etc. You should always log these fuckups. 

However, most of the time a *restart* of the bot is enough to keep things going.

<center><img src="https://media2.giphy.com/media/13FDZn3eBSzmNy/giphy.gif"></center>
<center><sup>Famous quote of <a href="http://www.imdb.com/title/tt0487831/">It Crowd</a>.</sup></center>

I use [Supervisor](http://supervisord.org) to control the processes of the bots. It monitors the processes and can auto-restart if something happens like that. It will also auto-start the scripts if an os reboot happens.

For example, I have [curobot](http://github.com/emre/curobot) as my personal curation bot. It's current config on my machine:

```
command = /var/bots/curobot-env/bin/curobot /var/bots/curobot/config.json
stdout_logfile = /var/logs/curobot.log
stderr_logfile = /var/logs/curobot.log
redirect_stderr = True
redirect_stdout = True
autorestart = True
environment = POSTING_KEY=foo
```

Supervisor daemon also has a neat web interface that, you can actually monitor and manage your processes via the web browser.

*A screenshot from my supervisord setup*

<img src="https://s13.postimg.org/m4wxwp5fr/Screen_Shot_2018-03-19_at_9.38.15_PM.png">

***

I use it for the last 4-5 years. It's a rock-solid solution to manage and monitor the processes. Check [supervisord.org](https://s13.postimg.org/m4wxwp5fr/Screen_Shot_2018-03-19_at_9.38.15_PM.png) for more info.