Sensor alert templates by device type
Noticed some new sensor alerts that popped up by default a few revs back now. Super sweet, firstly, so nice work guys. Secondly, the defaults naturally aren't quite right, ie 0. Is it possible to set a default sensor type by device/platform/OS rather than individually?
For example, system voltage on one of our routers has a default of 2 to 2.7v, while it is actually running at 23.5 and thus throwing alerts. Can we set all networking devices or iOS switches or iOS version 12.?? to have the same defaults?
Robbie Wright Siuslaw Broadband http://siuslawbroadband.com 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
Robbie,
If the device itself does not supply limits, Observium calculates limits (not in a very bright way, but still) based on the value of the sensor when it's first discovered.
Tom
On 27/01/2014 19:21, Robbie Wright wrote:
Noticed some new sensor alerts that popped up by default a few revs back now. Super sweet, firstly, so nice work guys. Secondly, the defaults naturally aren't quite right, ie 0. Is it possible to set a default sensor type by device/platform/OS rather than individually?
For example, system voltage on one of our routers has a default of 2 to 2.7v, while it is actually running at 23.5 and thus throwing alerts. Can we set all networking devices or iOS switches or iOS version 12.?? to have the same defaults?
Robbie Wright Siuslaw Broadband http://siuslawbroadband.com 541-902-5101
**For support issues, please email support@siuslawbroadband.com mailto:support@siuslawbroadband.com.**
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Use custom limits ;) http://awesomescreenshot.com/0d229j0z82
On Tue, Jan 28, 2014 at 12:26 AM, Tom Laermans tom.laermans@powersource.cxwrote:
Robbie,
If the device itself does not supply limits, Observium calculates limits (not in a very bright way, but still) based on the value of the sensor when it's first discovered.
Tom
On 27/01/2014 19:21, Robbie Wright wrote:
Noticed some new sensor alerts that popped up by default a few revs back now. Super sweet, firstly, so nice work guys. Secondly, the defaults naturally aren't quite right, ie 0. Is it possible to set a default sensor type by device/platform/OS rather than individually?
For example, system voltage on one of our routers has a default of 2 to 2.7v, while it is actually running at 23.5 and thus throwing alerts. Can we set all networking devices or iOS switches or iOS version 12.?? to have the same defaults?
Robbie Wright Siuslaw Broadband http://siuslawbroadband.com 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
observium mailing listobservium@observium.orghttp://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
On 27/01/2014 21:50, Mike Stupalov wrote:
Use custom limits ;) http://awesomescreenshot.com/0d229j0z82
On Tue, Jan 28, 2014 at 12:26 AM, Tom Laermans <tom.laermans@powersource.cx mailto:tom.laermans@powersource.cx> wrote:
Robbie, If the device itself does not supply limits, Observium calculates limits (not in a very bright way, but still) based on the value of the sensor when it's first discovered. Tom On 27/01/2014 19:21, Robbie Wright wrote:Noticed some new sensor alerts that popped up by default a few revs back now. Super sweet, firstly, so nice work guys. Secondly, the defaults naturally aren't quite right, ie 0. Is it possible to set a default sensor type by device/platform/OS rather than individually? For example, system voltage on one of our routers has a default of 2 to 2.7v, while it is actually running at 23.5 and thus throwing alerts. Can we set all networking devices or iOS switches or iOS version 12.?? to have the same defaults? Robbie Wright Siuslaw Broadband <http://siuslawbroadband.com> 541-902-5101 **For support issues, please email support@siuslawbroadband.com <mailto:support@siuslawbroadband.com>.** _______________________________________________ observium mailing list observium@observium.org <mailto:observium@observium.org> http://postman.memetic.org/cgi-bin/mailman/listinfo/observium_______________________________________________ observium mailing list observium@observium.org <mailto:observium@observium.org> http://postman.memetic.org/cgi-bin/mailman/listinfo/observium-- Mike Stupalov http://observium.org/
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
On 2014-01-27 15:08, Tom Laermans wrote:
Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex.
adam.
On 27/01/2014 22:51, Adam Armstrong wrote:
On 2014-01-27 15:08, Tom Laermans wrote:
Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex.
Agreed.
I would prefer smarter auto-limits. Like, for values surrounding 3.3v, use a static set of 3.0-3.5 or whatever, etc.
Tom
Agreed as well, I can see the complexity around templates as well. The only real complication was just bogus alerts and as Tom just mentioned, smarter auto-limits could be an easy way to fix that. They end user could then customize them individually if they needed something more, ie a switch in a non climate controlled cabinent on a telephone pole that ran hot.
Robbie Wright Siuslaw Broadband http://siuslawbroadband.com 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 2:57 PM, Tom Laermans tom.laermans@powersource.cxwrote:
On 27/01/2014 22:51, Adam Armstrong wrote:
On 2014-01-27 15:08, Tom Laermans wrote:
Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex.
Agreed.
I would prefer smarter auto-limits. Like, for values surrounding 3.3v, use a static set of 3.0-3.5 or whatever, etc.
Tom
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
On Cisco kit, the limits are provided by IOS as part of CISCO-ENTITY-SENSOR-MIB. If they're wrong, you should open a bug with Cisco.
As a workaround, we allow you to manually set the limits when Cisco fail.
adam.
On 2014-01-27 17:08, Robbie Wright wrote:
Agreed as well, I can see the complexity around templates as well. The only real complication was just bogus alerts and as Tom just mentioned, smarter auto-limits could be an easy way to fix that. They end user could then customize them individually if they needed something more, ie a switch in a non climate controlled cabinent on a telephone pole that ran hot.
Robbie Wright Siuslaw Broadband [1] 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 2:57 PM, Tom Laermans tom.laermans@powersource.cx wrote:
On 27/01/2014 22:51, Adam Armstrong wrote: On 2014-01-27 15:08, Tom Laermans wrote: Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex. Agreed.
I would prefer smarter auto-limits. Like, for values surrounding 3.3v, use a static set of 3.0-3.5 or whatever, etc.
Tom
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
Links:
[1] http://siuslawbroadband.com [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Cisco may have been a poor example. The problematic ones (with unsupported mibs) have been Mikrotik and the Ubiquiti wireless stuff. Ubiquiti's EdgeMax router is basically Vyatta so it picked up most things on its own, but not all.
Robbie Wright Siuslaw Broadband http://siuslawbroadband.com 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 3:25 PM, Adam Armstrong adama@memetic.org wrote:
On Cisco kit, the limits are provided by IOS as part of CISCO-ENTITY-SENSOR-MIB. If they're wrong, you should open a bug with Cisco.
As a workaround, we allow you to manually set the limits when Cisco fail.
adam.
On 2014-01-27 17:08, Robbie Wright wrote:
Agreed as well, I can see the complexity around templates as well. The only real complication was just bogus alerts and as Tom just mentioned, smarter auto-limits could be an easy way to fix that. They end user could then customize them individually if they needed something more, ie a switch in a non climate controlled cabinent on a telephone pole that ran hot.
Robbie Wright Siuslaw Broadband [1]
541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 2:57 PM, Tom Laermans tom.laermans@powersource.cx wrote:
On 27/01/2014 22:51, Adam Armstrong wrote: On 2014-01-27 15:08, Tom Laermans wrote: Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex. Agreed.
I would prefer smarter auto-limits. Like, for values surrounding 3.3v, use a static set of 3.0-3.5 or whatever, etc.
Tom
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
Links:
[1] http://siuslawbroadband.com [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
There are two ways the alert thresholds will be set :
a) From a manufacturer-supplied value, as with Cisco b) Based on the first measurement +/- a %age which varies based on unit (we allow temperature to vary more than voltage, for example)
The only way you'd have a sensor with a range of 2.0 - 2.5 when the measured value is ~23 is if :
a) The manufacturer supplied this range in the MIB b) The value at initial poll was ~2.3 c) The value passed at initial poll was an order of magnitude wrong, possible due to a bug in our code
There's no way for something to get a threshold an entire order of magnitude wrong without someone having broken something somewhere, either us or the manufacturer. If it's on our side, it's a bug that needs fixed, rather than something which needs worked around via more features.
adam.
On 2014-01-27 17:36, Robbie Wright wrote:
Cisco may have been a poor example. The problematic ones (with unsupported mibs) have been Mikrotik and the Ubiquiti wireless stuff. Ubiquiti's EdgeMax router is basically Vyatta so it picked up most things on its own, but not all.
Robbie Wright Siuslaw Broadband [3] 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 3:25 PM, Adam Armstrong adama@memetic.org wrote:
On Cisco kit, the limits are provided by IOS as part of CISCO-ENTITY-SENSOR-MIB. If they're wrong, you should open a bug with Cisco.
As a workaround, we allow you to manually set the limits when Cisco fail.
adam.
On 2014-01-27 17:08, Robbie Wright wrote:
Agreed as well, I can see the complexity around templates as well. The only real complication was just bogus alerts and as Tom just mentioned, smarter auto-limits could be an easy way to fix that. They end user could then customize them individually if they needed something more, ie a switch in a non climate controlled cabinent on a telephone pole that ran hot.
Robbie Wright Siuslaw Broadband [1]
541-902-5101 [1]
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 2:57 PM, Tom Laermans tom.laermans@powersource.cx wrote:
On 27/01/2014 22:51, Adam Armstrong wrote: On 2014-01-27 15:08, Tom Laermans wrote: Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex. Agreed.
I would prefer smarter auto-limits. Like, for values surrounding 3.3v, use a static set of 3.0-3.5 or whatever, etc.
Tom
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] [2]
Links:
[1] http://siuslawbroadband.com [3] [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
Links:
[1] tel:541-902-5101 [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3] http://siuslawbroadband.com
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
Thanks for the quick response Adam.
A is possible, but the Mikrotik MIB is unsupported (as of yet). I suppose the discovery service could have read some of the linux mib's and got confused since the actual MT mib isn't supported. Who knows. I'll just keep voting for full Mikrotik support. :-)
B Isn't likely as the device couldn't run on that voltage, but the board could have been giving off the wrong value anyway.
C No clue on this one, maybe just missing a decimal point. I guess we can see if anyone else has a similar issue or if I can reproduce it.
Robbie Wright Siuslaw Broadband http://siuslawbroadband.com 541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 3:43 PM, Adam Armstrong adama@memetic.org wrote:
There are two ways the alert thresholds will be set :
a) From a manufacturer-supplied value, as with Cisco b) Based on the first measurement +/- a %age which varies based on unit (we allow temperature to vary more than voltage, for example)
The only way you'd have a sensor with a range of 2.0 - 2.5 when the measured value is ~23 is if :
a) The manufacturer supplied this range in the MIB b) The value at initial poll was ~2.3 c) The value passed at initial poll was an order of magnitude wrong, possible due to a bug in our code
There's no way for something to get a threshold an entire order of magnitude wrong without someone having broken something somewhere, either us or the manufacturer. If it's on our side, it's a bug that needs fixed, rather than something which needs worked around via more features.
adam.
On 2014-01-27 17:36, Robbie Wright wrote:
Cisco may have been a poor example. The problematic ones (with unsupported mibs) have been Mikrotik and the Ubiquiti wireless stuff. Ubiquiti's EdgeMax router is basically Vyatta so it picked up most things on its own, but not all.
Robbie Wright Siuslaw Broadband [3]
541-902-5101
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 3:25 PM, Adam Armstrong adama@memetic.org wrote:
On Cisco kit, the limits are provided by IOS as part of CISCO-ENTITY-SENSOR-MIB. If they're wrong, you should open a bug with Cisco.
As a workaround, we allow you to manually set the limits when Cisco fail.
adam.
On 2014-01-27 17:08, Robbie Wright wrote:
Agreed as well, I can see the complexity around templates as well. The only real complication was just bogus alerts and as Tom just mentioned, smarter auto-limits could be an easy way to fix that. They end user could then customize them individually if they needed something more, ie a switch in a non climate controlled cabinent on a telephone pole that ran hot.
Robbie Wright Siuslaw Broadband [1]
541-902-5101 [1]
**For support issues, please email support@siuslawbroadband.com.**
On Mon, Jan 27, 2014 at 2:57 PM, Tom Laermans tom.laermans@powersource.cx wrote:
On 27/01/2014 22:51, Adam Armstrong wrote: On 2014-01-27 15:08, Tom Laermans wrote: Sure, but those are changed one on one, not with a template for all devices of a certain type as asked ;-)
I'm not even sure how you'd template these things, probably something like the alerting matching code, but i suspect it'd be little used, as it'd be really complex. Agreed.
I would prefer smarter auto-limits. Like, for values surrounding 3.3v, use a static set of 3.0-3.5 or whatever, etc.
Tom
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2] [2]
Links:
[1] http://siuslawbroadband.com [3] [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [2]
Links:
[1] tel:541-902-5101 [2] http://postman.memetic.org/cgi-bin/mailman/listinfo/observium [3] http://siuslawbroadband.com
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
On 28/01/2014 00:43, Adam Armstrong wrote:
There are two ways the alert thresholds will be set :
a) From a manufacturer-supplied value, as with Cisco b) Based on the first measurement +/- a %age which varies based on unit (we allow temperature to vary more than voltage, for example)
The only way you'd have a sensor with a range of 2.0 - 2.5 when the measured value is ~23 is if :
a) The manufacturer supplied this range in the MIB b) The value at initial poll was ~2.3 c) The value passed at initial poll was an order of magnitude wrong, possible due to a bug in our code
C was my initial thought, maybe we don't /10 $value upon discovery for some mib.
Tom
On 2014-01-27 18:13, Tom Laermans wrote:
On 28/01/2014 00:43, Adam Armstrong wrote: There are two ways the alert thresholds will be set :
a) From a manufacturer-supplied value, as with Cisco b) Based on the first measurement +/- a %age which varies based on unit (we allow temperature to vary more than voltage, for example)
The only way you'd have a sensor with a range of 2.0 - 2.5 when the measured value is ~23 is if :
a) The manufacturer supplied this range in the MIB b) The value at initial poll was ~2.3 c) The value passed at initial poll was an order of magnitude wrong, possible due to a bug in our code
C was my initial thought, maybe we don't /10 $value upon discovery for some mib.
Mine too.
adam.
participants (4)
-
Adam Armstrong -
Mike Stupalov -
Robbie Wright -
Tom Laermans