[Stork-dev] BIND9 stats not showing
Slawek Figiel
slawek at isc.org
Wed Feb 7 12:52:33 UTC 2024
Brendan,
unfortunately, I don't have good news for you.
Currently, there is no possibility to provide custom view name. The
"_default" value is hard-coded. I opened
https://gitlab.isc.org/isc-projects/stork/-/issues/1309 to change it and
make the view name configurable.
Additionally, I don't know a workaround for your problem with the
statistics endpoint. You are the first person who reports no statistics
if the "bind-dyndb-ldap" plugin is used. I have no experience with this
plugin, so I'm unsure if it is an expected behavior of BIND 9 or if your
endpoint is misconfigured. I opened
https://gitlab.isc.org/isc-projects/stork/-/issues/1310 to analyze this
problem.
You can also try to send a message to the BIND 9 mailing list. Maybe the
BIND 9 team or community would suggest a solution for you.
Regards,
Slawek
On 06/02/2024 22:52, brendan kearney wrote:
> the fact that i use views in my bind config is why there is no data
> showing up for one host. that host is a caching recursive resolver
> and a subordinate to my authoritative internal bind instances.
>
> on one of the authoritative bind instances, i am using views there
> too, which i might reconsider, i am using the bind-dyndb-ldap plugin
> to host the zone data in ldap, and when i curl for the json/v1 stats
> URL, only get back:
> [brendan at server3 ~]$ curl http://localhost:8053/json/v1/
> {
> "json-stats-version":"1.7",
> "boot-time":"2024-02-01T09:01:21.914Z",
> "config-time":"2024-02-01T09:01:27.425Z",
> "current-time":"2024-02-06T21:45:33.123Z",
> "version":"9.18.20"
> }
> even though the statistics are configured. it seems the caching
> mechanisms are not being used:
>
> ++ Cache Statistics ++
> [View: internal (Cache: internal)]
> 0 cache hits
> 0 cache misses
> 0 cache hits (from query)
> 0 cache misses (from query)
> 0 cache records deleted due to memory exhaustion
> 0 cache records deleted due to TTL expiration
> 0 covering nsec returned
> 0 cache database nodes
> 0 cache NSEC auxiliary database nodes
> 16 cache database hash buckets
> 5479735548 cache tree memory total
> 24906 cache tree memory in use
> 0 cache tree highest memory in use
> 243126656 cache heap memory total
> 1088 cache heap memory in use
> 0 cache heap highest memory in use
> [View: _bind (Cache: _bind)]
> 0 cache hits
> 0 cache misses
> 0 cache hits (from query)
> 0 cache misses (from query)
> 0 cache records deleted due to memory exhaustion
> 0 cache records deleted due to TTL expiration
> 0 covering nsec returned
> 0 cache database nodes
> 0 cache NSEC auxiliary database nodes
> 16 cache database hash buckets
> 24898 cache tree memory total
> 24898 cache tree memory in use
> 0 cache tree highest memory in use
> 1088 cache heap memory total
> 1088 cache heap memory in use
> 0 cache heap highest memory in use
>
> also, there is no _default zone, which might be due to my use of
> views, or different naming convention (_bind vs _default).
>
> is there a way to tell Stork that i want the stats from the views that
> i am running, or global stats from all views?
>
> On Tue, Feb 6, 2024 at 4:20 PM Slawek Figiel <slawek at isc.org> wrote:
>>
>> Brendan,
>>
>> > ugg... setting up a cottage industry of tools and services around
>> > stork for visualization is not what i was looking to get in to.
>> > especially since grafana requires prometheus. just too much to take
>> > on, for me and my one user... :D
>>
>> I see. Grafana and Prometheus tandem is easy to set up using Docker. But
>> I agree they are quite heavy components.
>>
>> Unfortunately, there are no plans to extend the visualization
>> capabilities in Stork. Please feel free to open a new ticket in the
>> Stork repository to describe your suggestions and address the problem.
>>
>> > in terms of the dropped packets, where do i look for what is being
>> > dropped? what is being dropped? traffic between stork server and
>> > agent, or something else? when i mouse over the icon, it says 0 hits
>> > and 0 misses. i am fairly certain that is not accurate. from the
>> > stats file:
>>
>> The query hit ratio is calculated using the data from the statistics
>> endpoint of BIND 9 ('/json/v1'). We retrieve the "QueryHits" and
>> "QueryMisses" metrics from the "_default" view of the returned data. The
>> ratio is calculated as "QueryHits / (QueryHits + QueryMisses)".
>>
>> If you see the zero values, it can caused by:
>>
>> - Communication problem between Stork server and agent
>> - Communication problem between Stork agent and BIND 9 daemon
>> - Not configured statistics endpoint in BIND 9
>> - No DNS traffic in your network
>>
>> Please, check your BIND 9 configuration and status of the machine and
>> application in Stork UI.
>>
>> Regards,
>> Slawek
>>
>> On 06/02/2024 22:07, brendan kearney wrote:
>>> ugg... setting up a cottage industry of tools and services around
>>> stork for visualization is not what i was looking to get in to.
>>> especially since grafana requires prometheus. just too much to take
>>> on, for me and my one user... :D
>>>
>>> in terms of the dropped packets, where do i look for what is being
>>> dropped? what is being dropped? traffic between stork server and
>>> agent, or something else? when i mouse over the icon, it says 0 hits
>>> and 0 misses. i am fairly certain that is not accurate. from the
>>> stats file:
>>>
>>> ++ Cache Statistics ++
>>> [View: bpk2 (Cache: bpk2)]
>>> 673135 cache hits
>>> 28 cache misses
>>> 191398 cache hits (from query)
>>> 162113 cache misses (from query)
>>> 0 cache records deleted due to memory exhaustion
>>> 137473 cache records deleted due to TTL expiration
>>> 3591 cache database nodes
>>> 2079 cache database hash buckets
>>> 3618520 cache tree memory total
>>> 1431595 cache tree memory in use
>>> 2595522 cache tree highest memory in use
>>> 524288 cache heap memory total
>>> 132096 cache heap memory in use
>>> 133120 cache heap highest memory in use
>>>
>>> thank you,
>>>
>>> brendan
>>>
>>> On Tue, Feb 6, 2024 at 3:48 PM Slawek Figiel <slawek at isc.org> wrote:
>>>>
>>>> Brendan,
>>>>
>>>> > the query hit ratio shows with a red exclamation point,
>>>> > which indicates that something is not being processed right.
>>>>
>>>> To be strict, the exclamation mark indicates the ratio is alarmingly
>>>> low. It means a lot of packets in your network are dropping.
>>>> It doesn't indicate any problems with processing data in Stork itself.
>>>>
>>>> > i thought data/visualizations like what is listed on this page
>>>> > (https://kb.isc.org/docs/monitoring-recommendations-for-bind-9) would
>>>> > be available.
>>>>
>>>> Unfortunately, Stork doesn't present more data in UI. But you can
>>>> integrate it with Grafana dashboard. Look at the screenshots under the
>>>> "Prometheus and Grafana" section on the main Stork repository page:
>>>> https://gitlab.isc.org/isc-projects/stork#prometheus-and-grafana
>>>>
>>>> We deliver some ready-to-use Grafana dashboards in the Stork server
>>>> binary package. One of them is dedicated to BIND 9. You can find them in
>>>> the "/var/lib/grafana/dashboards" directory.
>>>>
>>>> > is this something that may be seen in the future, or
>>>> > are things more "greenfield" at this point.
>>>>
>>>> Grafana is an excellent tool to visualize numeric statistics in time. We
>>>> decided to integrate with it instead of implementing similar
>>>> capabilities in Stork on our own. It allowed us to put more effort into
>>>> more complex Kea- and BIND 9-specific features.
>>>> The Stork agent acts like a Prometheus exporter. You can use it in this
>>>> role even without a Stork server.
>>>>
>>>> Regards,
>>>> Slawek
>>>>
>>>> On 06/02/2024 20:31, brendan kearney wrote:
>>>>> Slawek,
>>>>>
>>>>> maybe "dashboard" was not the right term to use. as you say, in the
>>>>> services page for bind9, i can see the bind instances on the servers
>>>>> that are running bind, and when i click on them, i can see limited
>>>>> details. the query hit ratio shows with a red exclamation point,
>>>>> which indicates that something is not being processed right. i
>>>>> thought data/visualizations like what is listed on this page
>>>>> (https://kb.isc.org/docs/monitoring-recommendations-for-bind-9) would
>>>>> be available. is this something that may be seen in the future, or
>>>>> are things more "greenfield" at this point.
>>>>>
>>>>> thanks for the quick response
>>>>>
>>>>> brendan
>>>>>
>>>>> On Tue, Feb 6, 2024 at 2:25 PM Slawek Figiel <slawek at isc.org> wrote:
>>>>>>
>>>>>> Hello Brendan!
>>>>>>
>>>>>> Unfortunately, no BIND 9-related data exists on the dashboard (except
>>>>>> for eventual events). The details of the BIND 9 daemons are available on
>>>>>> its specific application pages. You can list them by clicking the
>>>>>> "Services" button on the navbar and selecting "BIND 9 Apps".
>>>>>>
>>>>>> The current Stork server's capabilities to monitor BIND 9 are a bit
>>>>>> limited. We are open to suggestions on how to improve them. What data
>>>>>> would you like to have presented on the dashboard?
>>>>>>
>>>>>> Best regards,
>>>>>> Slawek
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 06/02/2024 20:17, brendan kearney wrote:
>>>>>>> list members,
>>>>>>>
>>>>>>> i have stork stood up and have the agent installed on a machine that
>>>>>>> is running bind9. the statistics channel and rndc control is
>>>>>>> configuured and working, but the stork dashboard does not show any
>>>>>>> stats for the bind instance. i can perform a curl for the statistics
>>>>>>> and get over 1 MB back in xml, so it seems that things should be
>>>>>>> working. i seem to be missing something, but am not sure what. how
>>>>>>> can i get the bind9 stats shown in the stork dashboard for bind
>>>>>>> instances?
>>>>>>>
>>>>>>> thanks in advance,
>>>>>>> brendan
>>>>>> --
>>>>>> stork-dev mailing list
>>>>>> stork-dev at lists.isc.org
>>>>>> https://lists.isc.org/mailman/listinfo/stork-dev
More information about the stork-dev
mailing list