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.
Comments
you need something with long-term historical data and alerts or rather on time metrics?
for the latter: https://www.netdata.cloud/agent/
Thank you. That's all I need.
Its great, but so much ram.
How much? Does it use ram only when page is opened. Or constantly?
Constantly in background around ~1-2 gb as i tried.
Please, not netdata! (Sending stuff out to a third party)
Munin or some other RRD monitor. Oldskool MRTG, even.
Examples (not mine):
https://cloud.addictivetips.com/wp-content/uploads/2019/04/Munin-Screenshot.png
https://programmi.giorgiotave.it/wp-content/uploads/2011/10/munin1-300x250.jpg
So, any recommendation?
I am just looking a simple htop. Shouldn't took much ram or cpu usage.
Just have! Munin: usually a very simple install (there's even a WHM/cPanel plugin, if that's what you have). I have it installed on webmin and CWP VPS servers too.
I've used it for decades and has helped to track down various issues, such as impending disc array failures, runaway processes (clamd) and apache/mysql memory allocation.
glances has a webui as well
Instal prometheus + grafana on a small vps, I have one installed on 386 MB Ram, then connect the client you would like to monitor using node-exporter, very efficient..
I just checked on one of my dedis and netdata uses just about 100k memory in background, I haven't looked very closely but I heavily doubt the said 1-2GB.
for seeing what kind of data it might send out and how to opt out, read here https://learn.netdata.cloud/docs/agent/anonymous-statistics
said that, I don't really share the concerns, but each to their own ;-)
It's the netdata Windoze-type attitude that really pisses me off. Opt out - sod that! It should be opt in. (My wifi has optout_nomap suffix, for any MS fans that may visit.)
Are you going to monitor the shell script for changes, just in case they (netdata) decide to also send URLs visited, for example?
My whole point is this though.. why not keep it all local/within your own infrastructure? Keep control over your own/customer's data.
There's Observium or its fork LibreNMS to consider too, both old school, but will monitor anything with a generic SNMP agent available, without running any extra proprietary plugins. Both fairly easy to set up and kick the tyres on if you have an underused VPS spare.
It could be disabled with --disable-telemetry, doesn't it?
I would suggest having a look at Zabbix https://www.zabbix.com/
Here at the company, we use Zabbix, Grafana and Munin.
Used to use this. It's ok: https://github.com/amonapp/amon
While I am annoyed to hear that this is an opt-out option, rather than opt-in,
AFAIK, netdata can be configured to export metrics to your private Prometheus aggregator. That should be a nice central place to query and graph against.
You can self-host any of the various configurations/features that HetrixTools offers. I’ve been using them for similar, albeit more ‘thorough’ requirements of my own across a sizable number of deployments spanning 9 different providers, without flaw for close to the last year or so now. So I don’t hesitate to recommend them.
observium is enough
Prometheus + Node Exporter + Grafana.
You only need to install Prometheus in just one server for data storage and Node Exporter in every server that need to be monitored.
For dashboard install Grafana in the same server as Prometheus or on your localhost. You can create your own dashboard or use the template like this https://grafana.com/grafana/dashboards/1860
May as well install docker, lavarel, composer, ansible, flatpak and any other software whilst you're at it. :-p
KISS