UsageTrend, our new feature that was introduced in the last RC for 3.2.0, has raised a number of questions:

The check is saying that the history store is not present. Where can I find it?

The -stm switch for the getter activates the history store setting. Of course, the corresponding object (aggregate|volume) has to be chosen as well. Please refer to the example below.

I am trying to understand the trend analysis:

./ -H $HOSTADDRESS$ ... --object aggregate --stm 100d

./ UsageTrend  -H $HOSTADDRESS$ -o aggregate –lookbehind=48h –full=80 -w 60d -c 30d

I waited for two days and somehow the check is still returning “UNKNOWN: Not enough history. You may want to review the checks –lookbehind and the getters –stm.”

With a -lookbehind of 48h I would suggest a -stm of 50h for the getter, but – 100d (days) are ok as well. One also has to pay attention to the -tolerance of the check. If it is smaller than the interval of the getter, it is quite possible that a corresponding call is found. Unfortunately, the default value in the last RC was much too small and it had to be set explicitely to a higher value (e.g 180 seconds) in order for it to work with a 3 minute interval of the getter:

$ ./ UsageTrend ... -o aggregate --lookbehind=48h --tolerance=180 --full=80 -w 60d -c 30d

NETAPP_PRO USAGETREND CRITICAL - 2 aggregates checked, 1 critical and 0 warning
p01 aggr1: 1899304.0 s (22.0 days) until full (80%), rate: 0.5 / day (CRITICAL)
p01 aggr0: 0.0 No change or negative rate in periode, rate: 0

Tip: If you want to verify the store contents of a trend check you can only do it through the check itself:

$ ./ UsageTrend -H sim821 -o aggregate --explore

Respectively, if you want to see only the timestamps of the calls (replaces the switch --explore=calls, that is currently unavailable for the explore of a check):

$ perl UsageTrend -H sim821 \
-o aggregate --explore | grep Call

------- Call from 1435744162 (2015-07-01 09:49:22 UTC)
------- Call from 1435743860 (2015-07-01 09:44:20 UTC)