Thu Jan 10 02:05:32 CET 2008
On Jan 9, 2008, at 4:02 PM, jm wrote:
> fess wrote:
>> I'm embarking on a project to monitor machines and internals of our
>> erlang applications using RRDTool.
> Can you elaborate on the project? Is it a stand alone monitor or
> something which could be rolled into one? I find our current
> system nagios somewhat limiting. As such I keep thinking about writing
> one, but like everyone else should be doing other things.
hmm, my goal is simply to make it easy to collect and view graphs of
I plan to be able to send "metric" messages from any part of our erlang
apps and have them stored to RRDs. [ how many widgets have I
how backlogged is my queue, how many errors am I ignoring ]
then have a graph browser tool and a graph designer tool, graph
designer tool would produce URLS to a REST style interface to RRD
the URLS would be placed wherever needed to make good presentations
about the state of the system [ wikis, other tools, etc. ]
I also plan to have a gateway for apps external to erlang to
feed metrics into the system, so we can add in regular
system metrics such as cpu/mem/disk usage, if we want everything
in one place.
Also, I'm not sure how I want to fit alerting off this data in,
but it would be an obvious next step.
So, it's kinda orthongal to nagios in that it's more about
metric collection, less about custom service monitors.
The big goal is that it become as _easy_ as possible to collect
and view data.
More information about the erlang-questions