Simple, free monitoring tool?

foxone

Hello,
I'm looking for a simple self-hosted monitoring tool that would let me see CPU, Network, Disk usage and similar stuff. An absolute requirement is that it needs to be self-contained and use low resources (gonna run it on LowEndSpirit instances).

So far i've tried:

Netdata

Pros:

• Very easy to install and use out of the box
• Self contained, no need for databases or other stuff
• Can aggregate multiple servers together

Cons:

• Not exactly "lightweight"
• Homepage is too confusing, it looked like a tutorial on how to read graphs rather than a simple dashboard
• The self installer pulls in a lot of garbage and pretty much takes over your entire system by putting files everywhere and being too much "magic" for my tastes

Monit

Pros:

• Very lightweight
• Only one binary file and a "/etc/monitrc" configuration
• Can very easily check running daemons and restart them / send alerts / do stuff automatically

Cons:

• There is barely a CPU meter, and no graphs.
• Aggregating multiple servers needs a license for M/Monit

Is there something else I should try?

Rahul

ops

SpeedBus

PHPSysInfo comes to mind, https://github.com/phpsysinfo/phpsysinfo

Demo: http://phpsysinfo.sourceforge.net/multi/index.php?disp=bootstrap

Neoon

Well, what I needed, was a simple monitoring, light weight, no dependencies, with notifications.
I tried netdata etc.. but I wanted to have something centralized, which sends me notifications and is a one liner to install.

So, I ended up writing my own, its will have a decent free tier of 30 servers. https://cryptic-sky.com
Its still alpha, has bugs, does not like OVZ..... but work in progress.

foxone

We let you know if someone is rendering porn at 3am.

Can i turn off these alerts?

mfs

Quick and dirty

$ pip install --user bottle glances
$ glances -w

apt install glances is fine too
et voilà, you have fancy graphs waiting for you at http://yiff.in.hell:61208/ (you obviously may put it behind a reverse proxy and so on)
Given you're talking about "aggregating multiple servers", Icinga and LibreNMS come to mind too, but none of those are the lightweight "fire and forget" solution it seems you're looking for tbqh

Neoon

@foxone said:

We let you know if someone is rendering porn at 3am.

Can i turn off these alerts?

Yeah.

terrorgen

netools

poisson

Maybe you can try this: https://www.monitorix.org/

Old school interface but definitely lightweight and can monitor unlimited servers.

comi

I have alerting system (not monitoring) assembled from shell scripts on cron with transport of choice (Telegram API in my case, but could be email or whatever)
I feel like if you want graphs it will never be truly lightweight. Though I've heard good things about LibreNMS.

freerangecloud

I use LibreNMS, the API is pretty good for automatically adding hosts from my IPAM solution and it supports Nagios plugins which means you can configure it to monitor just about anything. Plus it runs fine with 1GB of RAM and a single CPU core (currently monitoring 80 hosts, about 200 services)

Solaire

freerangecloud said: I use LibreNMS

+1 for LibreNMS, has been rock stable and it's very easy on resources.

turboprinz

I ended up with a simple bash script that does ssh to every server and checks resources. Simple and effective

tgl

@turboprinz i suggest to use ansible, its much better and it will basically do the same thing, there is nothing you need to setup on the servers

localhost

@turboprinz said:
I ended up with a simple bash script that does ssh to every server and checks resources. Simple and effective

Interesting. Can you share the script?

Rahul

localhost said: Interesting. Can you share the script?

I am pretty sure its bashed up beyond recognition.

skorous

Depending on how old school an interface you like, I'm actually using Xymon to monitor my LES nodes.

SpryServers_Tab

@SpeedBus said:
PHPSysInfo comes to mind, https://github.com/phpsysinfo/phpsysinfo

Demo: http://phpsysinfo.sourceforge.net/multi/index.php?disp=bootstrap

Yeah, that's what comes to mind for me too when he says his requirements.

They have a nice Android app too. (maybe iOS, not sure)

Janevski

@mfs said:
Quick and dirty

$ pip install --user bottle glances
$ glances -w

apt install glances is fine too
et voilà, you have fancy graphs waiting for you at http://yiff.in.hell:61208/ (you obviously may put it behind a reverse proxy and so on)
Given you're talking about "aggregating multiple servers", Icinga and LibreNMS come to mind too, but none of those are the lightweight "fire and forget" solution it seems you're looking for tbqh

This site can’t be reached

uhu

I just played with Nginx Amplify (not self-hosted), which does a reasonable job at monitoring nginx and system data on up to 5 "servers". Prometheus is probably better for self-hosting, with Chronograf or Grafana.

AlSwearengen

10 servers free with nodequery. Just not self hosted, which is a plus, IMO.