We manage about 50 different networks. One of the problems we have is they are all on the same private subnet behind a firewall (i.e. 10.1.1.0/24). Is there a way to have a central server where we could deploy "reporting" servers that poll the local networks for reports being sent to the central server? I hope that makes sense. If so, how would one go about accomplishing this?
-
-
If you can run a VM or Docker container, then a Zabbix Proxy server sounds ideal for your situation... -
We currently use solar winds, but it cannot handle identical subnets for different public IP's. So we do have monitoring systems at each location. From what you are saying, I am gathering that I can install a Zabbix proxy server and be able to accomplish what I am looking to do, correct?Comment
-
I haven't tried to monitor same subnets at multiple locations, but technically Zabbix can do it, as long as you have a globally unique identifier (eg. the IP/DNS) for the host and a Zabbix proxy at each location that will connect to the correct host when using the IP/DNS.
If you cannot detect a globally unique identifier, then Zabbix discovery (which scans an IP range) would be challenging/impossible to use.
In your case, you'll have run a Zabbix Proxy (with its associated DB) at each location and then run a central Zabbix server (with its associated DB) and a Zabbix web interface. The Zabbix Proxies will only act as a proxy/cache server and the Zabbix Server will collect all the data and do the alerting + reporting.
You can either let the Zabbix Proxy servers connect to the Zabbix server (aka Active mode) or your can let the Zabbix server connect to the proxy (aka Passive mode).
You can read more on Zabbix proxies here: https://www.zabbix.com/documentation...toring/proxies
So does your hosts/nodes have globally unique identifiers such as FQDN (even if it refers to a local private IP that also exists at a different location)?
What kind of API are you monitoring the hosts/nodes with? Just SNMP? Like Zabbix works fairly well for SNMP.Comment
-
Yes. Each location is actually behind a firewall with different public IP's. Each one is a separate domain. We are monitoring mainly SNMP for outages and latency issues. We need to also have notifications whenever a unit quits responding so we can be proactive in notifying customers of outages.
My supposition is that each proxy would be located behind the firewall for each subnet. Each proxy would be an active proxy as we need as close to real time data as possible. I would also assume we would then see each proxy server listed, and each node under that proxy then listed as a sub under it, correct?Comment
-
Yes, you'll need to have a Proxy server behind each firewall.
And yes, you can create a Host Group for each Proxy server and then group the Hosts that way. You can have Hosts assigned to multiple Host Groups too.
For instance we typically assign our Hosts to a Host Group for their location and a Host Group for the application name. This allow us to do aggregates for both the region and application, as well as ease searching for stuff that fits together
eg. Location/US East + Application/Customer Portal
It is probably not necessary to create a Host Group for each proxy, because you can also filter by the Proxy name and see its Hosts that its monitoring.Comment
Comment