thold 0.4.9 Re-Alert Cycle set to never but still Realerting

[DwayneDibbley]

running cacti 8.7h with thold 0.4.9

i have a threshold setup with a warning @ 15% and critical @ 0.01% the Re-Alert Cycle is set to NEVER, also disabled is the restoration email.

i had an alert on the 6th Jan reporting the threshold went below 15% with 14. but today i have received another alert reporting the threshold went below 15% with 4

the threshold is monitoring a printer toner supply level so should steadly decrease.

the graph also shows the level steadly going down

the question is why did thold send out another alert??

Thanks

[cigamit]

Good question, I will take a look at the re-alert logic to see if there is anything amiss. Also, are you using a CDEF or RPN to calculate the %?

[DwayneDibbley]

under the threshold template the following option is set

Under Data Manipulation, Data Type is Percentage and Percent Datasource is the supply_total. Also the min triggers for Warning and Alert are set to 1 hour ( 20x5min polls)

so far i am only getting the realerts for the warning thresholds.

Thanks

[DwayneDibbley]

I have now changed to actual values and still get the realert ( but is actually a new alert as no mention of still below, but instead went below )

i have looked at the device and it seems to get turned off overnight and the scenerio seems to be, afternoon alert triggered, overnight powered off so host down ( i have disabled host down alerts ), then following morning after being turned on the alert is resent ( new alert )

so it would seem that the thold data is erased or forgotten when the device is turned off and hence when turned back on it starts a fresh alert.

would this seem possible to disable the clearing when host down somehow?

[zpa0]

The same situation here, except that my host was always up.

cacti 0.8.7g
thold 0.4.9

Re-alerting happens every poller cycle, despite that Re-alert value is set to "Never".
This is true only for "Warning High / Low" settings, not for Alert Settings.