I have ArcGIS Monitor 10.7.1 monitoring several GIS' deployments. We have noticed that after a specific date & time, our 5 minute and 15 minute collections can randomly spike to take over 5 minutes. This is not too concerning for the 15 minute collection, but is clearly an issue for the 5 minute colleciton. Prior to this date/time, both collections were completing in under a minute. The date & time that this issue began on was a weekend and therefore an automated ICT issue is the likely cause (e.g. a windows patch rollout). Additionaly, when we receive the collection alerts, these align with a system counter that took a considerable time to collect and usually accounts for the increased collect time (e.g. a system collector for a singular VM took 320 seconds, 5 minute collection took 340 seconds). The VM that this system counter spikes/hangs on is random but it is always the system counter that is the underlying issue.
As Monitor was 100% performing as expected before this specific date & time, is there any common culprits we should be looking into that could degrade the system counter from behaving as normal? Or, any windows changes that could impact the counter? Once again, these spikes in collection time only occur a couple times a day and do not necessarily align with expect load on the system aka during business hours.