Today we had a an incident with a host in testing. It lost connection with vCenter, when we got it back up it was showing 100% cpu usage in vCenter but we couldn't find any correlation among the CPU dashboards we have. It was showing small spikes but mostly normal cpu usage. Nothing like what was showing in vCenter.
Prior to the host losing connection, vROPS was showing an alert stating there was a high number of anomalies.
So my question. What would cause this discrepancy among what I was seeing in vCenter vs. vROPS? My manager asked my opinion and I said it could be that some of the people who were involved in troubleshooting the incident didn't look at the right metrics. But after I dug around I couldn't see where the CPU had spiked on the Host in vROPS.
The only thing I can think of was that the CPU spike caused the connection to timeout with vCenter and when it came back it was showing high CPU. Since vROPS collects data at longer intervals it never saw it due to a timing issue...any thoughts?
