Pandora 2.0 could work nicely with 1000 agents or more (3.0 much more, give a try!).
LAG calculation is complex to explain here, I'll try:
You have a module of 300secs (5min) running.
Your agent has 7 minutes without any update (this means probably in that 2 mins agents is working gathering information, connecting to server and server is processing data). Real "lag" is 7-5 minutes, about 2 mins. This is more complex because you can have a lag in a module who is waiting for data (for example a event log module), and agent is not gathering it because cannot, so this module will count as "lag", and it isn't. Lag in data server is not very exact because the problems with unknown data and async modules. Data is more accurate for network/remote.
In the other way, 100% doesn't mean that servers are "on the limit", just sais that 100% of modules of that kind (data server) are processed by that server. It doesn't have a meaning of "processing limit capacity", just says that 100% of modules of that type are processed by that server, because you can have two or more servers of the same type running at the same time.
Using Pandora FMS 2.0 from VMWare Appliance download. I'm testing Pandora with a total of 7 Windows Agents, and in total 13 monitor checks. What's bothering me, is that the pandora_Data load always says it is working on 100% and there is always a LAG.
I want to install and monitor more Windows Agents, but will the LAG increase with more Agents?
I do have created a swapfile on the Ubuntu server, but that didn't work. When I take a look at the System Monitor, the CPU, Network and Memory Utilization is quite normal.
What to do to make it work with 25 Windows Agents?
