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.
Self-Hosted Server Monitoring?
Hello all,
I'd like to reduce my monitoring costs and am I'm currently (v happily) using both NodePing and Nixstats to monitor my setup. I have about ~22 servers, and this figure is likely to rise. I fully realise that neither of my current solutions are very expensive, but they are starting to add up.
Ideally, I'd like to find something that is self-hosted with NixStats' simplicity, but that doesn't have an insanely steep learning curve like Nagios. I tried Zabbix ~3-4 years ago and it was just too clunky.
What are you all using and would recommend?
PS: This is a great list of what's currently out there.
Comments
You won't find something like Nixstats that's self-hosted. If you are only concerned about ICMP/TCP monitoring then just get PhpServerMontior and put it on a cheap VPS. I use it to monitor dozens of servers and it's never failed me. It has Pushover, Email, SMS notification support.
Otherwise LibreNMS is the second thing I use for monitoring (up/down as well as resource monitoring.)
Then I use Nixstats on top of those two self-hosted/open source options.
I'm not, resource monitoring is probably more important than ICMP/TCP for me. I especially need to be able to see how many resources a particular process was using at a particular time last week for example.
How's it working for you?
I've used it for years, modern web interface, has all of the alert/notification support you would ever want.
Vigil is great. It's a kind of hybrid between a status page and monitoring, but you can configure alerts to Slack etc if any of your monitors go down.
It's made by one of the developers from Crisp IM and it's the exact same thing they use for the Crisp Statuspage.
Did you consider using something as "simple" as Munin?
Using it for years now and it's easy to setup and with the help of the right template no longer that harmful for the eyes. :P
Additionally, there are countless plugings available and writing your own one is very simple too. Specwise a small VPS should be more then enough for monitoring and if you set it up for on-demand generation if the charts is should be even lower on resource consumption.
@Bochi you got a screenshot of your template? :P
Nothing too spectacular, but also adapts well for mobile devices: MunStrap
Here is a example screenshot found in the repository:
Munin for simplicity
LibreNMS for better ui (I think)
Grafana + prometheus (including its exporter) for more complex and flexible monitoring
LibreNMS UI is good and easily to use
Have a look at statping
https://github.com/hunterlong/statping
Sounds like you need a central TSDB (time-series DB) for long term retention and querying.
Prometheus exporters (example: + Grafana would do.
If you need alerting, toss in AlertManager to this 'stack'.
IIRC you run netdata on each node for 'real-time view' and have it deliver metrics to your central Prometheus instance for 'long-term query/view'
Thank you, it was exactly what I was looking for, especially the remote Probe feature.
Thanks for this.
This looks awesome. Do you use it?
Does Prometheus collect the data and then Grafana render it? Then Alert Manager sits ontop to send alerts?
@alvin @akhfa LibreNMS looks good. Is it an Observium fork? I checked out the demo, is it possible to see/track how many resources individual processes are using?
Would anyone know of something similar to PHPServerMonitor, Statping, or Cabot that can also monitor SSL and DNS? Can LibreNMS do that?
Have a look at. https://www.zabbix.com/
It will zapp you hahahaha
I search PhpServerMontior in the Google , it's useful ,thank you
status2k
Yeah it's a fork of Obervium and yes it can monitor SSL through the Nagios plugin. Not sure about DNS.
Not yet, but will soon deploy and use it
https://my-netdata.io/
or
observium
Only these two. Both very different. Observium I do not like very much. Netdata can make lags on the server (confirmed several times, while devs telling cool stories that it does not affect a VPS, while it is).
I looked at it previously - it collects everything! All those graphs are a bit overwhelming
Thanks for pointing me towards the Nagios plugin.
I'm trying to figure it out if I can see how much RAM Apache (or any other process) was using last Tuesday with LibreNMS?
Blast from the past! Has this even been updated in the last few years?
@ehab said:
Trying to avoid behemoths like Zabbix, but thanks
Having seen the effects of netdata in CentosWebPanel, I'd say stay well clear. Just IMO.
Prometheus + Alertmanager + node_exporter. Grafana to graph it all.
It works for what it worth, simple.
>
nobody vote for Icinga?
Hmmm. Fork of Nagios, last release 19 March 2019. Interesting.
https://en.wikipedia.org/wiki/Icinga
currently im using it , if you want dashboard screenshot let me know
For my business i'm using the Check_MK monitoring.
https://mathias-kettner.com/
So you can install this backend on a vserver and deploy the agents on the server which you want to monitor.
The installation and the configuration is very easy
Zabbix has become very easy to setup and use in recent years. It is also not a resource hog anymore. You can monitor 300+ servers on a 4GB RAM VM.
I´ve been using Observium for some time and moved on to LibreNMS for maybe 3 years now... set it up on a cheap VPS... working fine measuring some 40 servers.. some of them really busy ones... All good with LibreNMS here...
I use elasticsearch and file/metricbeat on the client. It collects all metrics like cpu/hd etc and all logfiles are centralized. Works great for me, 25+ servers