@dch (assuming this message was supposed to cite me and is addressed to me)
I managed to fix the long-lasting trouble I had with syslog-ng and riemann leaking memory:
https://github.com/balabit/syslog-ng/pull/2523
Now, everything is smooth and I started doing relevant monitoring 😃.
riemann-dash happen to be a bit impractical, and I guess the next move for me will be to hack some kind of dashboard that better fits my needs (aggregating status by host / customer / etc).