Hi Pramod,
following are some figures (performance tests on a single CA SOI 3 Instance, 64-bit JVM):
- Event store captures incoming 1.000 events/second from connectors
- Manager processes 50 events/sec (modeled services) and 25 event/sec (event queues)
As Ashay mentions, SOI is not an Event Management system, but a Service Management system.
There are many processes running in background to evaluate the impact of Alerts towards Services, performing Escalations, assigning Alerts to queueus, keeping control of Access rights, etc.
Thus, the limiting factor of processing Alerts is not the Connector, but the complexity of the entire system, e.g. the SOI Manager.
The messages coming in will all be processed, but it might take a while until this is finished in case of a peek (Message Storm). This should never be the "normal" situation.
SOI is designed to handle "actionable" Alerts, e.g. messages that require an action (by an Operator or autmatically) such as opening a ticket, sending a mail.
No Operating Center can handle thousands of such Alerts on a regular basis.
You have to limit the information that comes into SOI (as Ashay says, best on Domain Manager level) to only forward to SOI what is really "actionable".
MichaelBoehm