Email notification for mrtg if over certain amount?
 

Any way to get notification if your traffic spikes over a certain amount? Ive hunted around before but never found anything to do this. Could be third party that checks every 20 mins or so. Anything like that?

i haven't seen a script that does it publicly available on the internet as of yet. we have it available to notify clients or us when bandwidth drops below or goes above a certain set amount though.

Use monit. MRTG is a pretty graph (and deprecated). Monit does shit.

Edit: BAM BITCHES. Welcome to 2005.

I was going to hit you up directly but couldnt find you in my list, I figured you knew. :winkwink:

Its hard to find in search engines cause you get so many home network and intranet monitors.

THEN PAY ME, BITCH!

I'm also 'off hours', e.g. "Leave me alone if you don't have the emergency number", e.g. "Willing to pay for off hour services."

Monit is fucking awesome. It works best when monitored OFF of the machine watched, but has worked a few times for me with a problematic single-server client to track issues. :thumbsup

I have more than one box running, so I could run it on the other box Im guessing? See if my host will install the bugger for me (as I am too lazy to do the homework on it hehe) if not I may hit you up later this week. I just want it to email me or whatever if bandwidth spikes over a certain amount.

As long as /server-status is setup correctly to allow that IP access, yes.. but it will not be able to stop/restart apache as it would for a local running daemon without security risks.

NOW IST THE TIME ON SPROCKETS VEN VE DANCE!

http://www.imagepup.com/up/5qFT_1210...-sprockets.jpg

haha.... God I miss the old Snl.. :Oh crap

So long as you have access to a text version of historical data, it shouldn't be hard to get grep or perl to spit out an email if usage exceed a certain percentage.
WG

Adding another process does so damn well when the machine's swapped itself to death.

It shouldn't be that consuming a process if its just scraping a dump of text, and it will probably just be run hourly.
WG

See, that's the problem. When the machine gets slashdotted, (or in my case, foobies'd several times when I ran blogs), it gets SLAMMED for hours, if not days. A local thing that fires off even every 5 minutes won't save it from a hard crash.

Im not too worried about a crash, just bandwidth going nuts for more than 20 minutes if I get a link to some chinese free site or something. Had some galleries get nailed a lot over the last year. Server does fine with 40mbps sustained for 45 minutes. Sites dont slow down or anything, but of course not good for the ole bandwidth bill. Had some shared pws too that I didnt catch in time. This is mainly like if my bandwidth spikes over 15mbps so I can take a look and see whats going on. Box hasnt crashed in a good 7 months since it got upgraded and moved to a new DC.

Are there screenshots anywhere? Or did those not exist in 2005?