19 Mar
2010
19 Mar
'10
1:53 p.m.
Hello,
I am wondering if anyone has a clever way to remotely monitor a Kamailio 1.5
server. I am not looking for the standard monitoring, what I am looking to
achieve is catching situations where my upstream carrier is having
problems. We have a certain level of 404, 500, 503 errors throughout the
day which are not indicative of a major carrier problem. I want to be able
to monitor the ratio of properly setup calls to failed setups - so I can
know when a carrier is having issues and is responding with many 503 errors.
Any push in the right direction would be greatly appreciated.
thanks.
19 Mar
19 Mar
2:28 p.m.
You can log negative replies from a failure route, or put them in a
database, or issue an HTTP request.
On 03/19/2010 01:53 PM, Geoffrey Mina wrote:
Hello,
I am wondering if anyone has a clever way to remotely monitor a Kamailio
1.5 server. I am not looking for the standard monitoring, what I am
looking to achieve is catching situations where my upstream carrier is
having problems. We have a certain level of 404, 500, 503 errors
throughout the day which are not indicative of a major carrier problem.
I want to be able to monitor the ratio of properly setup calls to failed
setups - so I can know when a carrier is having issues and is responding
with many 503 errors.
Any push in the right direction would be greatly appreciated.
thanks.
_______________________________________________
Kamailio (OpenSER) - Users mailing list
Users(a)lists.kamailio.org
http://lists.kamailio.org/cgi-bin/mailman/listinfo/users
http://lists.openser-project.org/cgi-bin/mailman/listinfo/users
--
Alex Balashov - Principal
Evariste Systems LLC
Tel : +1 678-954-0670
Direct : +1 678-954-0671
Web : http://www.evaristesys.com/
21 Mar
21 Mar
11:11 a.m.
Hello,
On 3/19/10 7:28 PM, Alex Balashov wrote:
You can log negative replies from a failure route, or
put them in a
database, or issue an HTTP request.
On 03/19/2010 01:53 PM, Geoffrey Mina wrote:
> Hello,
> I am wondering if anyone has a clever way to remotely monitor a Kamailio
> 1.5 server. I am not looking for the standard monitoring, what I am
> looking to achieve is catching situations where my upstream carrier is
> having problems. We have a certain level of 404, 500, 503 errors
> throughout the day which are not indicative of a major carrier problem.
> I want to be able to monitor the ratio of properly setup calls to failed
> setups - so I can know when a carrier is having issues and is responding
> with many 503 errors.
>
> Any push in the right direction would be greatly appreciated.
>
if you have a fixed number of carriers and use snmp, then you can define
some statistics.
I use htable with 3.0 and dump them via mi/rpc command. I do not need to
have persistence, otherwise I would use database.
Usually at the beginning of production I catch every reply in
onreply_route and have classes of replies per carrier in hashtable, kind of
if(status=~"4[0-9][0-9]")
{
$sht(ht=>$si::4xx) = $sht(ht=>$si::4xx) + 1;
}
Dumping the content via rpc from time to time for analysis, which is
good for checking after restarts.
Cheers,
Daniel
--
Daniel-Constantin Mierla
* http://www.asipto.com/
22 Mar
22 Mar
1:32 p.m.
On Friday 19 March 2010, Geoffrey Mina wrote:
I am wondering if anyone has a clever way to remotely
monitor a Kamailio
1.5 server. I am not looking for the standard monitoring, what I am
looking to achieve is catching situations where my upstream carrier is
having problems. We have a certain level of 404, 500, 503 errors
throughout the day which are not indicative of a major carrier problem. I
want to be able to monitor the ratio of properly setup calls to failed
setups - so I can know when a carrier is having issues and is responding
with many 503 errors.
Any push in the right direction would be greatly appreciated.
Hi Geoffrey,
you probably do this already, but i've found the ASR rate (200 OK - failures)
which we generate from our accounting a really helpful value to quickly notice
a QoS problem for a certain carrier. Especially if you can compare against
past data, like the last day, or last week. If you then add NER (200 Ok -
"carrier failures") it gives you an even better picture.
You probably should be able to build something like this also with the htable
approach Daniel suggested and some web graph tool.
Cheers,
Henning
5447
days inactive
5450
days old
3 comments
4 participants
participants (4)
-
Alex Balashov -
Daniel-Constantin Mierla -
Geoffrey Mina -
Henning Westerholt