up vote 0 down vote favorite
share [g+] share [fb]

I'm trying to find a tool that will:

  • monitor all of my devices and services on < 10 linux servers,
  • allow me to view graphs of historical data - not moving averages, but real stats historically,
  • configure alerting depending on different thresholds for each of the monitored services
  • be easy to extend so I can monitor custom metrics/servers - this must be easy like with nagios/ganglia.
  • be as easy as possible to deploy in that it will auto-discover services available and begin collecting standard data straightaway.
  • is fully self-contained and doesn't require a load of plugins to do anything useful and doesn't require me to check multiple interfaces. I'm looking for a single software package.

I've been trying Hyperic, and it works except that it's a pain to extend with custom monitors for really simple stats gathering.

I'm considering zabbix or zenoss. zabbix looks better, but does it satisfy all of my criteria? How about zenoss? Is there anything else which is ideally free/reasonably priced that would fit the bill?

Thanks

link|improve this question
feedback

6 Answers

up vote 1 down vote

Zenoss Core is my go-to favorite.

link|improve this answer
feedback
up vote 0 down vote

Maybe you should separate the 2 task you have at hand.

  • is real time monitoring of systems

For the first one there is commonly zabbix and nagios. I have not heard of anything that is freely available.

  • monitoring of history

Well - history is logs. Thus, you need a log management system. There are front-ends to syslog-ng that i found easy to set up and use years ago - but i think you should google more - lot's changes in 3 years.

I've also recently found this prodcut - http://www.splunk.com/ - but can't comment on it.

link|improve this answer
but why should i have to push the same metrics in two places because most monitoring tools won't preserve data? i don't need it preserved indefinitely. ~6 months should do so i can spot trends. unfortunately ganglia - which i like - uses moving averages which means that while it keeps data for a year, spikes get averaged out over time making it difficult to spot ceilings, etc. I looked at splunk earlier though and it looks good, but i don't think it's for me right now. – smith Nov 16 '10 at 16:04
Why not? It's free for up to 500 hosts if i remember correctly. They also have a nice set of plug ins. Thу interface also looks fantastic. All in all it's definitely a solution to consider. Futhermore there are allways add-ons for monitoring systems that are often not even related to the project owner. Maybe that's the direction you should be searching in. I doubt that there will be an obscure monitoring system that will suite your needs exactly. – Redconnection Nov 17 '10 at 2:25
feedback
up vote 0 down vote

Cacti... :D

it managed for a clients 25 servers through,

  • snmp and service pings..
  • Simple Notifications and thresholds
  • Graphing using rrd's
  • Also through another plugin, showed the syslog server

If your interested i can tell you more..

Hope this helps :D

link|improve this answer
feedback
up vote 0 down vote

Take a look at Zabbix.

For every value you want to check, you can:

  • set triggers/thresold
  • set the number of days of your data
  • generate graph
link|improve this answer
feedback
up vote 0 down vote

I'd look at OpenNMS and Epylog, personally.

Both are 100% free in their entirety - no "community" vs "commercial" editions.

link|improve this answer
feedback
up vote 0 down vote

I would check out Nagios XI it's awesome, it does cost money but not much in comparison to allot of companies. I had to pay a one time fee of 1300 bucks and there supporting me for a year and giving me upgrades for free. and I was able to get help from ethan him self that was pretty cool to me.

link|improve this answer
feedback

Your Answer

 
or
required, but never shown
discard

Not the answer you're looking for? Browse other questions tagged or ask your own question.