Hi all,
I agree with Jeroen. Until we don't develop a "LPIC monitoring" we can
avoid getting too tied to a particular product and just know a bit about
them, but, in any case, these are current
http://popcon.debian.org/by_inst [3] figures for several monitoring
solutions:
* PRODUCT --- by_inst --- by_recent
* Nagios: 4303 5081
* Icinga2: 11628 10157
* Zabbix: 4873 5767
* Prometheus: 28236 24125
* Shinken: 53086 150600
* Sensu: 23046 15921
So, clearly, Nagios, Zabbix and Shinken are going down whilst Sensu,
Icinga2 and Prometheus go up. Among the 3, Icinga2 is the most used, but
it's still behind Nagios and Zabbix.
I've deployed several Icinga2 with many timeseries backends, but
everyone speaks nice about Prometheus, its builtin timeseries and how
fast is it on a single setup and how well does it adapt to
"services-philosophy".
Zabbix is simply crappy stuff with a silly design to store timeseries
into a RDBMS.
IMHO this is a good topic where trainers can show a bit about their
preferred monitoring system without getting too much into details. Just
mention some of the top opensource solutions and leave it as-is until we
create a more specific course or increase the scope of this to show how
to perform a minimal setup in 2-3 of these platforms (like we do with
FTP servers on LPIC2 or VPN servers in LPIC 303).
Otherwise, my vote goes for Icinga2 :-)
Regards,
Kenneth
A 2017-07-06 16:51, Jeroen Baten escrigué:
> Hi all,
>
> Well, you can teach a dog a trick, but humans are able to understand things.
>
> This means that the question is not "icinga or prometheus" but
> understanding monitoring.
>
> And having knowledge of how to setup monitoring.
>
> A short google trends shows a higher result on Prometheus but this can
> also be attributed to a certain movie.
>
> And, why not Zabbix now we are talking about this? Easy to install from
> a repo and anyone with a browser can configure it. Just click away like
> any self respecting Windows admin. (yes, this is a subtle rant :-) )
>
> So, all the descriptions are non-specific regarding a product.
>
> I would suggest to make the exam also non-specific. After all this has
> always been one onf the strong points for LPI. Distribution agnostic
> certification. Why change that now?
>
> - Just my 2 cents
> - your humble cervant
> - more self-humiliation to avoid starting a flame war....
>
> Warm regards,
>
> Jeroen Baten
>
> Op 06-07-17 om 15:25 schreef Fabian Thorns:
>
>> Dear all,
>>
>> you probably know we're getting closer to the finalization of the
>> LPIC-OT DevOps Tools Engineer certification. While we're doing some last
>> tweaks to the objectives, we struggled over the question whether Icinga
>> 2 or Prometheus is the 'better' tool for monitoring the components of an
>> application. Also, from a didactic point of view, Prometheus seems to be
>> easier to learn than Icinga.
>>
>> What are your thoughts about replacing Icinga 2 with Prometheus in here:
>>
>> https://wiki.lpi.org/wiki/LPIC-OT_DevOps_Tools_Engineer_Objectives_V1#705.1_IT_Operations_and_Monitoring_.28weight:_4.29
>> [1]
>>
>> Thanks for your thoughts,
>>
>> Fabian
>>
>> _______________________________________________
>> lpi-examdev mailing list
>> [email protected]
>> http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev [2]
Links:
------
[1]
https://wiki.lpi.org/wiki/LPIC-OT_DevOps_Tools_Engineer_Objectives_V1#705.1_IT_Operations_and_Monitoring_.28weight:_4.29
[2] http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
[3] http://popcon.debian.org/by_inst
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev
