In internal.lists.toasters, you wrote:
Does anyone track or log the load (CPU, NFS ops, network in/out, etc.) on their Network Appliance boxes (over time)? If so, I'd like to hear about your experiences.
We track everything we can get from sysstat and netstat. It gets written to log files and timestamped where we can graph it on-demand via a CGI script and form.
We've found it to be a very helpful aide in diagnosing problems. If we start to see sluggish performance, we can at least determine if something is different today than it was two weeks ago.
We are also able to use this setup to predict when it is time to order another filer, disks, NIC, etc.
Tripp