Advice on improving stability

[paul]

Hi,

I'm reasonably new to the zimbra world, and have the zimbra community edition running.
I've been having occasional problems with stability, and I'm looking for advice on
improving it.

I'm very nervous of updates as the most recent couple took down the server.
One was caused by a file being set to read only. I dont know what the other one was, but
running the zmfixperms sorted that out.
Each of these took a couple of hours of researching, experimentation, scanning logs etc.

More recently, after an automatic overnight reboot, mail wouldn't flow. I didn't notice until
the morning. There's a troubleshooting document here:
https://wiki.zimbra.com/wiki/Incoming_Mail_Problems
which says that if /var/log/zimbra.log is empty (mine always is), its likely DNS or firewall issues. I confirmed that
it was neither of those. (BTW - that document is very short and not very helpful. Is there something
better?). All /opt/zimbra/log files looked fine, 'zmcontrol status' showed everything running. In desperation,
I tried another reboot and everything started flowing again.

I've also had the issue of zmstat filling the hard drive. I've fixed that thanks to
viewtopic.php?t=65930&start=10
However, any upcoming update could undo the fix to zmstat-io, and fill the hard drive again.
That concerns me so I've stopped it.

To my questions:
a) is there a better troubleshooting guide that could walk me through symptoms and solutions?
b) is this level of service failure normal? I'm coming from postfix, dovecot, and later with roundcube and owncloud where
the server ran for 6 flawless years.
c) what tips do you have to improve service reliability?

Thanks for any advice.
Paul