Need little input on a solution design.
We have around 90+ small/tiny hosts ( by capacity) scattered across the globe and requirement is to monitor them. These are all 1/2 CPU, 2 Gig memory, 8 - 10 GB physical storage boxes and we installed Hub software on them to avoid security issues ( by opening 10 ports to public).
Now the problem is as there are too tiny hub itself consuming 90-100% CPU not leaving any computation power for the service that the host intended to run and as result we are getting hundreds of Queue disconnection and service degradation alarms to the console.
Apart from hub we are also running CDM, processes and net_connect which are almost the killing the host.
Now, need little help in taking the hub software off those hosts but still monitoring basic parameters like CDM, processes etc.
Note: Turning them into robots is a tried out option and we paused it as it is opening too many number of ports to the internet and package deployment failing despite critical ports being opened.