The whole idea of hosting AHM on Amazon EC2 was to avoid situation when the site is down and cannot send out the message that there is something happening. I was under impression that RMA agents provide exactly the same functionality in terms of tests and "just" report to AHM "server". Server itself can check if the site(s) running RMA agents is down and then report that there is an issue.
Not sure what exactly means "site" in this context. Server where HostMonitor is started? Server where RMA started? Entire LAN?
> avoid situation when the site is down and cannot send out the message >that there is something happening
For each RMA agent you may setup backup RMA agent.
To monitor HostMonitor itself, you may use WatchDog, another instance of HostMonitor or may be Amazon CloudWatch (I am not sure how exactly it works).
If you already using Amazon EC2, just download and install HostMonitor, you need 5 min to check if it works. Well, it should work because you are running Windows Server 2008. Another question how exactly implemented this Amazon cloud-cluster-failover thechnology. We don't know details, but I assume it should be reliable...
If you already paying many for EC2, why not use it?
Two of the sites are 100 and so users, servers and network stuff, third site is just few users.
Not sure I understand how "users" can be related to monitoring configuration. HostMonitor monitors hardware and software.
Do you mean there are 100 HostMonitor operators that need to control monitoring in some way using remote access to HostMonitor (RCC or Web Service)?
or you mean HostMonitor should check 100 workstations plus some servers?
IF RMA is limited for whole site monitoring, maybe we should consider two licenses for AHM for each of the largest sites that will monitor itself and other site using HM Monitor feature and later adding the Cluster controller should we need it?
It depends on your needs. e.g. If you need consolidated reports and logs (eg. one report for all systems located in 3 networks), then you need 1 HostMonitor instance + agents.
Otherwise you may setup several instances of HostMonitor and control all instances from one location using several copies of RCC. E.g. each HostMonitor may monitor its own LAN plus check another HostMonitor and some critical parameters in another LAN.
Regards
Alex