What do you suggest to use for cPanel/DirectAdmin server monitoring?

someshzade

What do you suggest to use for cPanel/DirectAdmin server monitoring,

I checked Nagios and found its complex for me to configure it.

AlwaysSkint

Munin

K4Y5

HetrixTools?

I'd rather not waste time in setting up my own monitoring system / agents and then setting up another set of agents to monitor my primary monitoring agents. Life is too short for that nonsense.

LeonDynamic

@K4Y5 said:
HetrixTools?

I'd rather not waste time in setting up my own monitoring system / agents and then setting up another set of agents to monitor my primary monitoring agents. Life is too short for that nonsense.

Absolutely agree. Go with Hetrixtools, you can monitor everything from there, setup different notifications to whatever platform you want and away you go.

AlwaysSkint

All depends on how much information that you want and how much you wish to expose to external parties. The OP was lacking in detail, apart from Nagios. Munin is an appropriate alternative for internal use.

96.9%

uptime, according to HetrixTools for a one server which I know is nonsense.

seriesn

@AlwaysSkint said:
All depends on how much information that you want and how much you wish to expose to external parties. The OP was lacking in detail, apart from Nagios. Munin is an appropriate alternative for internal use.

96.9%

uptime, according to HetrixTools for a one server which I know is nonsense.

If you run out of disk space, HT will show you offline if you are using Agent-based monitoring.

AlwaysSkint

HT can show you offline, even without agents, for http(s),mysql etc.
With agents/snmp much more is possible.
Perhaps I should've emphasised my first sentence in the previous post.

Daniel

Netdata is really awesome. Here's a post I made about it on another forum: https://hostballs.com/t/netdata-awesome-system-monitoring-tool/1662

It has per-second granularity for all the metrics it collects, so can be extremely useful for debugging CPU usage spikes that only last a few seconds (as one of the metrics it collects is CPU and RAM usage per process, and per systemd unit). A lot of other monitoring systems are only per-minute so they don't catch small spikes.

AlwaysSkint

Wouldn't go anywhere near the netdata spyware, after an encounter in CWP.

someshzade

Thanks to everyone, much appreciate your suggestions.

Rahul

@someshzade

So, which did you pick ?

Mr_Tom

@AlwaysSkint said: Wouldn't go anywhere near the netdata spyware, after an encounter in CWP

What's this in reference to? Is there a link/etc to read more?

AlwaysSkint

@Mr_Tom said:
What's this in reference to? Is there a link/etc to read more?

In CWP, click on Graphs, Netdata and before you know it, it installs software and sends a heap of data to a 3rd party. Fab. not!
The act of monitoring implies some performance degradation, so for 1s granularity I suggest the monitoring itself is a consumer of resources. Want to see an example of this in action? Try Windows Task Manager with its' stupid default refresh rate, then decrease it to 5sec. More noticeable on lower CPUs and it only displays a few metrics but the fundamentals are the same.

Mr_Tom

@AlwaysSkint said: In CWP, click on Graphs, Netdata and before you know it, it installs software and sends a heap of data to a 3rd party

Ah so it's a CWP issue, rather than netdata itself?

someshzade

@Rahul said:
@someshzade

So, which did you pick ?

Kinda confused now.... but taking advice from other experts as well. If you have any suggestion please drop me a DM.

AlwaysSkint

@Mr_Tom said: Ah so it's a CWP issue, rather than netdata itself?

Yes, though the type & quantity of data passed to the remote netdata was far too revealing and was enough for me to immediately terminate it. Fine to use external montoring (Hetrix, Uptime etc.) to check for common ports (eg. 80,ping) but best to keep detailed server stats for internal consumption. You could run your own central monitor, if you feel the need for consolidation - I had a freebie VPS doing this duty, for 6 months; only for testing, as it's better than idling, just.
Note: I am biased having come from an Enterprise Management background.

Daniel

@AlwaysSkint said: The act of monitoring implies some performance degradation, so for 1s granularity I suggest the monitoring itself is a consumer of resources.

I've installed Netdata on several servers and have never seen it use more than 5% CPU. I've even got it running on NAT VPSes, although I decreased the refresh rate to 5 seconds on those (not due to CPU usage, but due to RAM usage). For me, CPU usage of Netdata collecting metrics once per second is actually less than Munin was when I used it to collect metrics once per minute. The core of Netdata is written in C which contributes to its lean-ness.

@AlwaysSkint said: In CWP, click on Graphs, Netdata and before you know it, it installs software and sends a heap of data to a 3rd party

This sounds like a CWP thing (whatever that is)... Default netdata doesn't send data to a third party.

AlwaysSkint

Might be tempted to take another look at netdata, installing from github. Munin can sometimes get a bit bogged down generating charts but generally I've found it reliable.

Rahul

@someshzade said: If you have any suggestion

nodequery.com works