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.
Nagios to monitor 7 Servers, and alternatives to ?
Hello I manage 7 servers and i always used a basic tool to monitor cpu load, apache, mariadb, etc.... to give me an alert if something goes wrong. Looking for something, finally, more interesting like Nagios® Core and i'd like to understand if a basic vps with 1 or 2 dedicated cores (with a very very good uptime, no bandwidth limit) is ok for this work.
Do you have other suggestions for a tool for this job ? (p.s is not a problem if setup requires some work)
Comments
Monit + apprise. Have Monit perform passive monitoring + forwarding. For events beyond your control send to a centralized mail account and use apprise to deliver a suitable push notification. Don't desensitize yourself to the data, but focus on the anomalies.
You could use Nagios, Icinga, CheckMK / OMD etc.
I currently use Icinga, paired with a bot which sends me messages via Telegram if something goes wrong. You don't really need much ressources to get a reliable working setup with 7 client hosts.
We worked with https://bloonix-monitoring.org/. But have now switched to https://www.cloudradar.io/.
Both tools are highly recommended by us.
I'm using Nagios to monitor like 30 servers, works great.
Running it on a € 2,99 OVH sandbox VPS with very low specs.
What I like about Nagios are the many custom monitoring scripts you can download for it. I can't speak about the others, never used them.
https://www.librenms.org
Prometheus, AlertManager and Grafana with node_exporter put onto all servers. The modern way of doing things (also the best)
Just a snippet of what you can see, there's metrics for everything imaginable.
Zabbix. It can be deployed on Raspberry PI.
I fully endorse this setup, I implemented it to replace Zabbix and while I absolutely loved Zabbix my infrastructure has never been this easy to monitor.
Glad you like it. When we switched, life became a lot easier. We manage pretty much everything with Ansible so when we want a new Prometheus exporter on a server, I implement or create an Ansible role that installs, updates and manages the configuration for it on each server. To maintain it our monitoring we just have to add, remove or change a few variables in our playbook vars. I'm surprised people still use these old monitoring tools like Nagios, infrastructure as code is the way to go!
If anyone is interested in what our Ansible playbook looks like to do this, check it out! https://git.ulayer.net/snippets/52/raw
Any setup guides?
Looks awesome
We might make a blog post in the future, but it'll be about how to deploy this kind of setup with Ansible.
How does netdata compare to Prometheus+AlertManager+Grafana?
Been hearing a lot about Prometheus and Grafana lately.
I'm keen to try a pure Python monitoring stack of some kind but haven't figured out what. Probably Graphite + something.
Netdata is a local collection daemon AFAIK. It can sub in for node_exporter and export collected samples to prometheus.
Prometheus is primarily pull-oriented though by design.
AlertManager is a prom plugin IIRC
Grafana is a 3rd party graphing solution.
So The prometheus 'stack' is a giant pile of Legos; a.k.a Build your own monitoring DeathStar battle group.
Endorse Prometheus and Grafana with bunch of exporters, we are heavy users of grafna, we have build many panels [in react] and data sources [node.js], also using many buildin data sources and panels comes with Grafana.