New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
What do you use for monitoring (self-managed, not a service)?
raindog308
Administrator, Veteran
Looking for:
- monitoring public services (http, etc. - the usual)
- monitoring private info (disk space, memory, etc.)
- alert me when needed
- ideally something whose agent is very memory-light and can run on 128MB LEBs
- agent runs on Debian, Ubuntu, CentOS
- the master I'm less sensitive about but should fit inside a 512MB VPS
- free...or very low cost
- about 10 servers total to monitor
I was going to setup nagios for monitoring, but
- the plugins package doesn't compile after following the install guide
- the docs say they are incomplete (and they are). People say it's complex to setup but the docs are weak...that's what I've heard. I'd rather not embark on a voyage of discovery.
- I'm annoyed to discover there were packages I could have apt-getted instead of following the official docs (which build from source)...why the docs don't just say "run apt-get" is beyond me.
So - what else?
- Munin, though I don't think monitoring is really its strength or focus
- Zabbix - heard very nice things about this
- Observium, though again doesn't seem like a monitoring product per se
- Monit - haven't read much about this
Comments
OpenStatus
Openstatus. Does a good job of monitoring disk space, memory, cpu, bandwidth, and services. Shoots me an email when I need it to. Light on the resource usage as well.
http://openstatus.nickmoeck.com/
I'm currently using Zabbix on a couple servers. It has lots of templates for various services, like generic linux or mysql or memcached, etc. It also has configurable alerts. DotDeb hosts the latest versions for debian, so easy apt-get install.
I can't speak to memory use, as I have it on non-LEBs and configured to be high-performing, not low resource use. I know that you can tweak the config to spawn fewer processes.
We use Zabbix monitoring a few thousand servers and it scales quite nicely. Their enterprise support and custom development services are great.
We were long time nagios/icinga users prior and man... Zabbix is sooooo much better.
Munin is perfect for me. Easy to set alerts, easy to do what I've done with the graphs on my blog. I run it for 3 nodes and munin is hosted on ramnode 128mb.
I've been toying with observium and I do like it so far, but not sure I'd like to take the time to perfect an snmp setup.
Running zabbix agent on a 32mb LEB debian. Using the version from the zabbix repos.
zabbix 2338 0.0 0.0 6268 4 ? S Feb22 0:00 /usr/sbin/zabbix_agentd
zabbix 2339 0.0 1.0 6268 352 ? S Feb22 3:14 /usr/sbin/zabbix_agentd
zabbix 2340 0.0 2.2 6284 728 ? S Feb22 2:15 /usr/sbin/zabbix_agentd
zabbix 2341 0.0 2.2 6284 752 ? S Feb22 2:15 /usr/sbin/zabbix_agentd
zabbix 2342 0.0 2.2 6284 744 ? S Feb22 2:15 /usr/sbin/zabbix_agentd
zabbix 2343 0.0 1.0 6268 332 ? S Feb22 0:30 /usr/sbin/zabbix_agentd
munin+icinga
@tchen - running a monitoring agent on a 32mb LEB warms my LowEndHeart.
monit.
you can use munin (with cgi graphics and html on demand) on a larger node to monitor all your vps, but if you need something small to monitor, alert and restart daemons go with monit. Really easy to setup, light and uses very little memory. Give it a try.
I use munin+monit+ an esternal script i wrote to monitor webpages content and services that alerts me via SMS when there's a serious problem.
@NickM
I'm having problem with OpenStatus, installed it on debian6 with your repo, configured server up, configured client, made nginx vhost. I access the page, log into admin panel and try to add server - it does not get added to the list.
I restarted php5-fpm, made sure I had all the requirements.
Debugging the server it only shows receiving JSON from the client.
Do you have any idea what's up? I have no idea. Could it be some permission problem, or nginx?
Only errors I receive are from nginx vhost error log:
I personally prefer CopperEgg.com they have a very nice interface, free trial check them out
Started with Cacti.
Started using Munin.
Stopped using Cacti
Started using Observium.
Still using Munin and Observium. Munin makes it easier to monitor anything that's not accessible via SNMP.
https://github.com/mojeda/ServerStatus
Thought about writing a cron job that would ping me if any of the servers are down, but haven't done so yet. While I do like programs that you can just install and use, there is a certain satisfaction you get from doing it yourself (ish).
Also using UptimeRobot.com to email me at @vtext.com (Verizon USA) to SMS me.
I do this too. A lot of uptime services offer "SMS credit" to send messages, but most (all?) cell carriers have an email address that will send the contents of the messages to the cell phone. And they're free (other than standard text msg costs, if any)
Using Mojeda's though I was hoping for something with IN/OUT statistics, think I'll have to go the Munin route or OS.
Munin and Nagios. But sometime on small project this is overkill and I just use hosted service called StatusCake.
I'm using too many I think, depending of several factors (not all controlled by me) I'm using:
nagios
opennms
zabbix
zenoss
munin
:-(
@prometeus if you had the chance to winnow that list down to two, which would it be?
without thinking about it I wrote the list in order of preference :-)
so here you go: nagios and opennms or zabbix
Monitoring 10 servers would be $90/month. Too much.
Thanks for the replies everyone. I think I will check out Zabbix and OpenStatus.
Collectd - very lightweight
Does using IRC to see if servers ping timeout count?
N2...free, open source, written in c and c++, highly scalable, repos available for debian/ubuntu and centos/fedora or compile from source
http://opensource.cloudvps.com/n2/
opsview open source
If anyone wants to make a lil' bit o' cash - setup Icinga for me.
zabbix for hardware monitoring + site24x7 for uptime monitoring
I've just switched over to StatusCake - statuscake.com. It's free and I haven't had any problems with them yet.
Well, freemium...free + pay for better features. But looks interesting. Thanks.
Not a bad idea!
which is easier to setup?