Feature request: max item value (integer)

[timmy]

Situation: UPS temp sensor reporting 4 Giga Celsius Temperature (max for integer type variable) when external temp sensor is not attached to UPS.
Result: unreadable graph due to scale - 4G C' and 20-23 C' in 'day' timeline.
Suggestion: Create field to limit value got from item when creating it. Maybe it'll be usefull for some devices with snmp query results we cant control...

Comments?

[richlv]

i'd suggest logging a bugreport with ups vendor to fix the firmware...
for now maybe external check/userparameter that filters out such values could be used

[timmy]

UPS output

Just in case:

UPS model: Smart-UPS RT 5000 XL, External temp sensor supported but not attached.
Returned temperature value: -1 (in Zabbix it looks like max value of INT type)
I tried to use "Throw map" to make "0" from "-1" but graphs still unreadable when UPS returns -1.

[richlv]

oh, if it's returning -1 and zabbix records max possible, that's not good. please report this on the bugtracker

[den_crane]

Situation: UPS temp sensor reporting 4 Giga Celsius Temperature (max for integer type variable) when external temp sensor is not attached to UPS.
Result: unreadable graph due to scale - 4G C' and 20-23 C' in 'day' timeline.
Suggestion: Create field to limit value got from item when creating it. Maybe it'll be usefull for some devices with snmp query results we cant control...

Comments?

just disable this item on this host.


APCProbeCurrentTemp
.1.3.6.1.4.1.318.1.1.10.2.3.2.1.4.1

APCBatteryTemperature
.1.3.6.1.4.1.318.1.1.1.2.2.2.0

[timmy]

Cant do

Those corrupted temp values got occasionaly by the UPS (rarely). Most of the time UPS reports normal values thus the item is useful.

PS: Sometimes limit of the item is needed for other devices - for example cluster of 2 network devices on the edge of the network: provider bandwith = 10Mbit/s. When quering cluster for in/out traffic we can get max 9.7-10 Mbit/s, but when one cluster node initiates a failover traffic values can reach 150-300 Mbit (on 1G link) - this affects badly on the graphs.
(Devices 2xJuniper ISG-2000, NSRP Active/Passive)

[den_crane]

Those corrupted temp values got occasionaly by the UPS (rarely). Most of the time UPS reports normal values thus the item is useful.

change item value type from INT to FLOAT. Try use snmp v2 instead of snmp v1

[timmy]

change item value type from INT to FLOAT. Try use snmp v2 instead of snmp v1

Yep. 'Float' type returns -1. Graphs look good.