I want to build a script which would check up an instance of RabbitMQ on my server once in a minute. Is it possible to check if RabbitMQ is "doing well" in automatic mode, it might be via a script (ruby, python, whatever) or command line. By "doing well" it's not about to crash for any reason and it's not frozen.
Also, if I'm able to connect to it from a client script, say, from ruby, does that mean that "it's doing well" or not necessarily?
That doesn't mean it's doing well. The problem is that "doing well" cannot be measured. You need to check things like total queued messages, messages per second or memory consumption. A simple ping won't tell you much. Heck, rabbitmq as an erlang system is built to crash and respawn.
Once you define what you mean by doing well, you can create a script to hit Rabbit's API. It's simple HTTP.
The API becomes available via the Rabbit Management Plug see https://www.rabbitmq.com/management.html Once installed you have to define what it means to be doing well within the context of your application. It could be that you app takes a long time to process messages. It could be that you'll have bursts of messages but you have to average their processing time. It could be that you purposely under powered the server so you only want to worry about extreme memory pressures. See http://looselycoupledlabs.com/2014/08/monitoring-rabbitmq/ for an example metric set.
The is no single stats value that will tell you any server is about to fail. You'll want to combine Rabbit's stats with the host OS'.