Hello team, I’d like to understand Harvest capabilities to monitor memory utilization outside of EMS events. I see we expose node_memory but that’s total memory rather than utilization.
Any help would be greatly appreciated!
#OOM Node condition monitoring
nimble vale
sweet escarp
I might be wrong but I don’t think ONTAP advertises statistics regarding detailed memory utilization. Memory is usually used, that’s a good thing, you wouldn’t expect to have lots of free memory for a “healthy” system
nimble vale
We’re in a situation where Wafl ran out of memory but didn’t trigger a failover until 45 minutes later. Aside from EMS, there’s no way to monitor for this condition? We’d like to preempt if a node is heading in the wrong direction so we can alert and failover manually.