We're getting alerts about our MDS 9509s utilizing high memory; the problem is there is actually little free memory but a ton of cached memory meaning that the switch is actually healthy.
Example:
sh sys resources
Load average: 1 minute: 0.00 5 minutes: 0.02 15 minutes: 0.00
Processes : 177 total, 1 running
CPU states : nan% user, nan% kernel, nan% idle
Memory usage: 1033100K total, 981236K used, 51864K free
10044K buffers, 419960K cache
Is there any way to get NPM to take this into account or to adjust the threshold for individual nodes?
Thanks,
Mitch