Hello Dan,
There are several things in your message that I wanted to comment on.
"it was close to a week before any one noticed it was down,"
Do you mean that the users did not notice that the WAAS was not active?
"because its backup failed"
It is difficult to image a scenario where a WAAS in a pair would not take over the traffic from a failing member.
How do you do traffic redirection and why did the backup not work?
To answer your concern:
- to know if the WAE is overloaded we need to check for the overload alarms, but also for prolonged periods of high CPU or excessive cache turnaround, so a low maximum cache age. You can check for both these things from the GUI.
- to measure of a WAE is having an effect the best way is to get objective numbers. Doing a timed copy from the command line is often the best way, as for the GUI is often misleading in reporting timings.
It can be that the WAAS is optimizing traffic but the users in general do not notice, sometimes because there is a network problem that slows down the traffic again (duplex mismatches often are the cause for this) or something as mundane as that the improvement is measurable but not perceived by the end users.
To end: WAAS in general does not need tuning, the one thing most customer do is to exclude traffic that cannot be optimized or is misbehaving.
Best regards, Peter