Which network monitoring system to choose for 250+ hosts (windows, linux servers, routers, switch)?

T

Tremo2016-01-20 05:23:40

linux

Tremo, 2016-01-20 05:23:40

Good morning to all, and good evening to some.
This is not the first time I have contacted the toaster with questions about the monitoring system for my office. I have already read a lot of things and learned about how it should work, thanks to which some kind of representation of the monitoring system began to emerge.
I read a lot and have already managed to experiment with zabiks, which in my case cannot send a request to devices. He successfully manages to monitor only the zabbix server.
Comrades, I'll tell you a little about my network and dare to ask you a couple of questions.
There are 250+ hosts on the network which include Linux servers, windows servers, routers, switches.
There is a main server machine on which everything should be displayed.
The goal is to install a monitoring system on the host machine. The system must support templates, graphics, SNMP, traps, alarms, and be relatively robust. Preferably open source or some not very expensive solution.
As far as I understand, the scheme is as follows. SNMP manager is installed on the main machine from which it will send requests to devices on which the SNMP agent should be installed.
Any hints and advice will be glad.

Reply

Answer the question

In order to leave comments, you need to log in

8 answer(s)

I

Ivan, 2016-01-20
@t3mp

Zabbix

C

civdef, 2016-01-20
@civdef

I suggest to start to see what is simpler. For example dude from mikrotik. Normal ping (icmp v4) without agent installation. When you enable SNMP on hosts, you can freely receive data from all sensors. A bunch of material from the how-to series. + Easy to write queries. If you don't like it look for something else.

V

Vincent, 2016-01-20
@Vincen_t

nagios

S

sillik, 2016-01-20
@sillik

Hyperic HQ
habrahabr.ru/post/24854

C

Cool Admin, 2016-01-20
@ifaustrue

If the network is predominantly Windows, you can look towards the System Center product family.

K

Konstantin, 2016-01-20
@fallen8rwtf

at the old job there was double monitoring:
the windows structure was monitored via PRTG (they now give 100 sensors for free), it is very convenient, quickly deployed and at least in
Linux configuration, switches and other snmp crap - via zabbix. She's perfect for that, I think.
did you open ports for receiving? I suspect that this is the problem. for localhost it is open, but for an external adapter, judging by your problem, it is closed.
there was a structure with printers (monitored all the parameters on them, up to the number of scans in duplex - it was necessary), hp (3com) and cisco network equipment, Linux servers. it took about a month to perfect it (I was the only admin) + the channel with the remote office was monitored

A

athacker, 2016-01-20
@athacker

We monitor infrastructure on Zabbix:
Number of hosts (enabled/disabled/templates) 5765 5512 / 111 / 142
Number of items (enabled/disabled/not supported) 538807 523119 / 12809 / 94
Number of triggers (enabled/disabled [problem/ ok]) 130488 128402 / 2086 [17 / 128385]
Templates, graphics, SNMP, IPMI, traps - it can do all this.

T

Talik, 2016-01-21
@Talik0507

Definitely Zabbix. At first glance, a huge incomprehensible thing, but upon closer examination, it is a very simple, convenient and functional thing. Expandable functionality limited only by your imagination