On 02/10/2016 08:50 PM, George Peristerakis wrote:
> I have 2 suggestions:
> 
> 1. After looking at log-pusher, I would suggest a alternative, 
> http://www.rsyslog.com/. It can be configured in the jenkins-slaves.
> 2. Logstash is  good, and I think this video [1] is useful to understand how 
> others are designing log aggregation on top of openstack.  
> 
> [1]. https://youtu.be/1ye0-sityBw

Rsyslog and the video you suggested seems to be designed for *system*
logs...

What we want here is ci jobs' logs such as jenkins console output.
The goal is to give SF users a mean to search for a specifics content
through all jobs, to measure the number of occurrence or identify which
projects are affected by a new random bug for example.

So the challenge is to collect slave ephemeral logs and add some
contents (such as the related gerrit change-id and job name) so that the
database lookup yield more meaningful data.

Perhaps Rsyslog or Fluentd could do the jobs, but it seems like the
openstack infra is relying on logstash and a custom script to collect
the logs:
http://docs.openstack.org/infra/system-config/logstash.html



Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Softwarefactory-dev mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/softwarefactory-dev

Reply via email to