Thanks to everyone that replied.
Haven't decided which way to go yet, but thanks for all the good
info!
Just to clarify a bit, I'm not putting the Pis under any kind of
heavy load. They will be running in environmentally "unfriendly"
areas (garage, patio, etc), so I want to ensure that they are not
overheating, especially in summer (I already have them in custom
cooled cases).
thanks!
Peter
On 10/30/2021 6:18 PM, Matt Graham via
PLUG-discuss wrote:
On
2021-10-29 17:57, AZ Pete via PLUG-discuss wrote:
3 Raspberry Pis, 1 Ubuntu, 1 Linux Mint,
and am looking for a way that I
can monitor them remotely.
You probably don't need to do this unless they're under much
heavier load than is usual for home machines.
a unified dashboard that would list for
each server:
CPU utilization history, Drive space consumed, Temperature
monitoring
(CPU, chip set, & hard drive temps), Fan speed monitoring,
Some kind
of alerting mechanism when a given threshold is passed
I've looked into Webmin, Glances, Nagios,
Collectd & Cockpit. Mostly
these seemed to be geared more toward remote admin. But the
monitoring
that was available didn't seem to include the temperature info
or the
idea of one unified dashboard.
When I was doing this (a long time ago), it was not possible
without 2 separate programs. We used Ganglia to keep records of
CPU load, disk space, free RAM, number of database connections (if
the machine was a MySQL server), and other stats. Ganglia's
default configuration had a web page that showed various
statistics for all the machines that are set up and running the
client Ganglia service.
Ganglia does not notify people about things though. To send mail
to people or put a notice on a web page that said, "WARNING:
machine foo-1234 has more than 200 active database connections",
we used Nagios. The Nagios server can monitor any parameter that
is measurable from the Nagios clients, because Nagios monitors are
(were?) Perl scripts that run on the clients. Perl can easily
parse the output from `sensors` or `df` or `free` and return "OK",
"Warning", or "Error" as you wish.
There used to be a rather useful firefox extension called
nagios-checker that would poll a Nagios web page and display
useful information in firefox's status bar. However, they have
improved firefox so much that the extension no longer works.
If installing and configuring this stuff sounds like too much work
for a tiny number of machines that are probably not under very
heavy load, you're right. It might be a useful learning
experience, but its *practical* value is probably not very high.