On Thu, Sep 12, 2002 at 08:39:10AM -0700, Jack Coates wrote: > On Wed, 2002-09-11 at 10:02, Oliver Thieke wrote: > > > > Finally I MUST write some custom monitors (checks) and this > > seems to be easier (and better documented) in MON. > > Looks like you've already summed it up pretty nicely. A co-worker of > mine is messing with Nagios right now, and has been dismayed to find > that it's got HTML hard-coded into some of the C modules. This makes it > pretty difficult to modify the output screens (for instance, to link to > MRTG from a Nagios screen).
I think we can sum it up like this: Get a pretty monitoring package if you need a "For Dummies" setup that simply tells you if really common services, which are set up in really common configurations - are working. Get Mon if you need a monitoring framework. One where you can: - write your own monitors - write your own alerts - define different time intervals for different checks - feed alerts to another Mon server or a typical NMS via traps - set dependencies on alerts - avoid cluttering your servers with weird monitoring clients - or don't even bother writing your own stuff, use the contributed stuff but know that you can customize your monitoring at a moments notice if you suddenly need to So that's the summary, in my humble opinion. If you need a framework, Mon can't be beat. If you need pretty and dumb, get Big Brother or something like that. Don't get me wrong, BB works if BB is all you need, it's just that serious sites outgrow it before they even get started. -- "Training is everything. The peach was once a bitter almond; cauliflower is nothing but cabbage with a college education." - Samuel Clemens
msg00644/pgp00000.pgp
Description: PGP signature
