We have several hundred Linux VMs on EC2 and Google compute engine. We want to monitor basic things like disk free space and memory consumption, in the easiest and lightest way possible. Expectedly, VMs come and go pretty often, as load changes, etc.

Currently we use simple scripts that pull such information via SNMP. We don't need fancy app-specific monitoring since it is already being provided by app-specific means.

We tried Zenoss, and found it hard to use, and its documentation lacking.

We considered Nagios and its forks. We considered Sensu (but my boss is not a fan of RabbitMQ) and Ganglia, but all of them seem a bit too complicated for our very basic needs.

SaaS solutions like Circonus would be too expensive with the number of hosts we have.

Am I missing some obvious easy solution here? What would you recommend [against]?

closed as primarily opinion-based by xhienne, Scott, Stephen Rauch, GAD3R, Kusalananda Sep 9 '17 at 15:48

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    Why aren't you happy with your SNMP solution? – Red Cricket Feb 27 '13 at 17:59
  • Many of those you mentioned can poll using SMTP parameters. – mdpc Feb 27 '13 at 18:26
  • @RedCricket: mostly because it's homegrown, requires changing code to modify, doesn't readily show pretty graphs, etc. This can be added, but why reinvent the wheel? I'm pretty sure I'm not the first person with a need like this, but I failed to find a simple-to-deploy basic monitoring solution. This seems improbable; I ought to have a blind spot in my google-fu. – 9000 Feb 27 '13 at 18:35
  • We use Nagios here, but I wouldn't recommend it for your application. AFAIK, Nagios doesn't work so well in highly dynamic environments. – derobert Feb 27 '13 at 22:00

If you are looking more into the open-source-direction, Open NMS might suit your needs. I did not use it myself, but I heard good things about it (especially from people, who dislike Nagios). From what I read about it, it is also SNMP-based.

  • Seems to do the work. – 9000 Mar 2 '13 at 22:52

I would think that collectd might fit your needs, especially in EC2, as you can add it to your base AMI, and it will send data to (one or many) central collector(s).

That way new instances immediately start sending data back to you, and you're not always adjusting your monitoring system with add/removes.

Use graphite to collect the data, and a simple nagios setup or cron job script to check for values that are out of the 'norm.'

From that base you can grow easily.


My personal favorite is WhatsUp Gold Premium.

Low cost and based on SNMP with very good (historical) graphing abilities. On mouseover it will show a split-second actual reading.

Only drawback is that it needs a Windows-Server and a MS-SQL-Database (it comes with the free light edition).

It is a commercial solution for small environments like yours.

  • I haven't tried your solution because we're a windows-shy shop, but I do appreciate your calling our setup 'small' :) – 9000 Mar 2 '13 at 22:53
  • @9000 Is "medium" better?-) Small was with regards to monitoring. If you start to monitor routers each port can become a monitoring-object - that will very soon be big (with regards to the number of monitored devices). – Nils Mar 4 '13 at 12:22

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