OpsCX integration capacity, OEM, Nagios, Zabbix etc
Is there a documented maximum capacity when a third party tool has a certain number of nodes and objects?
What is the maximum capacity of a single integration to third party tools such as Zabbix, OEM and Nagios?
Specifically, at what point will Events, Topology or Metrics fail or have errors if the number of objects in the third party tool has reached a certain point? Events, metrics and topology are brought in serially.
I suspect I may have hit the max with Zabbix. The node count is over 2,000 in Zabbix. Topology takes over an hour to bring in. I looked at metrics and it doesn't appear it can process the entire node list of metrics in 5 minutes.
There is no actual limit to the number of nodes OMi can monitor.
When you are configuring OMi you have to set the "Server Deployment" where you indicate the amounths of nodes OMi will monitor, being the smallest deployment of 2000 nodes. OMi can monitor even more than 5000 nodes but you will have to assign more resources to OMi with higer numbers of nodes.
You can check the details of the memory resources for the deployments at the installation guide of OMi, at "Check the hardware requirements".
My question is on the capacity of OpsCX, not OMi. There must be a limit on how much data OpsCX is able to integrate for events, metrics and topology.
For instance, if OpsCX is integrated to a third party system that has 10k nodes. Try to pull in 20 metrics per node every 5 minutes (for PMI graphing). Can OpsCX keep up? Pulling through API calls 200K of metrics every 5 minutes?
I would think not, but I don't see anything published on the subject and how one would scale the integration with Medium to Enterprise level third party systems.