Different polling frequency for certain graphs
We are using the standard polling frequency of 5min, which is fine for normal (customer) interfaces. But regarding core interface (in our MPLS network) 5 min polling type does not show short outages etc.
Is there any way to poll certain graphs (e.g. limited number with a certain tag) with another frequency (e.g. 1 min), Or do I have to have a another observium instance for that?
Regards MKS
1 min won't show short outages either.
Don't start chasing the magic dragon.
Adam.
Sent from BlueMail
On 13 Mar 2017, 14:02, at 14:02, MKS rekordmeister@gmail.com wrote:
We are using the standard polling frequency of 5min, which is fine for normal (customer) interfaces. But regarding core interface (in our MPLS network) 5 min polling type does not show short outages etc.
Is there any way to poll certain graphs (e.g. limited number with a certain tag) with another frequency (e.g. 1 min), Or do I have to have a another observium instance for that?
Regards MKS
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
That's what SNMP traps are for!
We are using the standard polling frequency of 5min, which is fine for normal (customer) interfaces. But regarding core interface (in our MPLS network) 5 min polling type does not show short outages etc.
Is there any way to poll certain graphs (e.g. limited number with a certain tag) with another frequency (e.g. 1 min), Or do I have to have a another observium instance for that?
Regards MKS
Sure, but e.g. changing metric on core links and having two wait 10 min to see the (propper) effect (two polling cycles) v.s 2 min when the polling cycle is 1 min. That is a big difference. And I really not need to poll customers interfaces with 1 min, 5 min is fine for that.
On Mon, Mar 13, 2017 at 3:10 PM, Michiel Klaver michiel@klaver.it wrote:
That's what SNMP traps are for!
We are using the standard polling frequency of 5min, which is fine for
normal (customer) interfaces.
But regarding core interface (in our MPLS network) 5 min polling type
does not show short outages etc.
Is there any way to poll certain graphs (e.g. limited number with a
certain tag) with another frequency (e.g. 1 min),
Or do I have to have a another observium instance for that?
Regards MKS
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
We use the realtime graphs under those scenarios.
You can get a nice 10sec polling (or less if your device sports it) graph to see immediate changes.
Michael
On 25 March 2017 2:40:34 AM LHDT, MKS rekordmeister@gmail.com wrote:
Sure, but e.g. changing metric on core links and having two wait 10 min to see the (propper) effect (two polling cycles) v.s 2 min when the polling cycle is 1 min. That is a big difference. And I really not need to poll customers interfaces with 1 min, 5 min is fine for that.
On Mon, Mar 13, 2017 at 3:10 PM, Michiel Klaver michiel@klaver.it wrote:
That's what SNMP traps are for!
We are using the standard polling frequency of 5min, which is fine
for
normal (customer) interfaces.
But regarding core interface (in our MPLS network) 5 min polling
type
does not show short outages etc.
Is there any way to poll certain graphs (e.g. limited number with a
certain tag) with another frequency (e.g. 1 min),
Or do I have to have a another observium instance for that?
Regards MKS
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
participants (4)
-
Adam Armstrong -
Michael -
Michiel Klaver -
MKS