Thanks for reply.
I am not 100% sure that this was a resource issue...
In the case Rohit has mentioned, the full system restart has fixed the problem and all nodes are available now.
I did a single node restart first (the one that was offline), but it did not help.
Full system restart (all 7 nodes) have fixed the issue.
My main concern is how to detect this problem in the future.
We are running a services status check on all nodes and it sends the alert if the critical service is down.
In this case all services were running and I did not notice that we have a problem until I've looked at QMC over the weekend.
Apparently, the disconnect happened on Friday...
Any suggestions how we can detect this kind of issues in the future?