On 05/20/2016 11:29 AM, Jason DeTiberus wrote:
On Fri, May 20, 2016 at 1:22 PM, Rich Megginson <[email protected]
<mailto:[email protected]>> wrote:
On 05/20/2016 10:54 AM, Jason DeTiberus wrote:
On Fri, May 20, 2016 at 10:40 AM, Rich Megginson
<[email protected] <mailto:[email protected]>> wrote:
We are trying to start up a CentOS OpsTools SIG
https://wiki.centos.org/SpecialInterestGroup for logging,
monitoring, etc.
It almost seems to me that this is actually a meta SIG in a
sense. I would almost expect there to be a SIG for each sub topic
here.
So we should have a logging SIG, a monitoring SIG, etc.?
I believe so. I'm not completely against them living in an OpsTools
SIG, I just worry about OpsTools focusing on one type of logging
and/or monitoring framework and then a competing SIG coming along that
would address others.
Well, for logging, we will be initially focusing on fluentd, rsyslog,
elasticsearch, and kibana, since these are the most popular, and are
already used today by OpenShift Origin and RDO. If someone wants to
support logstash, I wouldn't be against that. Would/does this overlap
with any existing SIGs?
The intention is that this would be the upstream for
development and packaging of tools related to logging (EFK
stack, etc.), monitoring, and other opstools, as a single
place where packages can be consumed by OpenShift Origin,
RDO, and other upstream projects that use CentOS - pool our
resources, share the lessons learned, and enable cross
project log aggregation and correlation (e.g. running
OpenShift on top of OpenStack on top of Ceph/Gluster - do my
OpenShift application errors correlate with Nova errors?
file system errors?).
I definitely love the concept, I just want to make sure that we
don't duplicate effort being done by the existing SIGs or end up
with conflicting efforts.
This would also be a place for installers (puppet manifests,
ansible playbooks), and possibly testing/CI and containers.
So, for OpenShift we already have the PaaS SIG that will cover
installation and testing/CI. The Cloud SIG covers this for
OpenStack as well.
What about testing/CI for running OpenShift with an integrated EFK
stack? Would that be covered by the PaaS SIG? Same with Cloud
SIG, running OpenStack with an EFK stack for logging.
I believe we would want to extend that into the PaaS SIG (I can't
really speak to the Cloud SIG), since logging and metrics are an
integral part of the complete OpenShift platform. Obviously we need to
do work towards the automated deployment of those platforms, but I
would fully intend that testing and CI coverage include the deployed
logging and metrics components. Shipping containers would also have to
be tied in closely with the PaaS SIG, since platform versions are tied
to versions of the integrated containers as well.
Can we use the Pass SIG testing/CI framework for testing OpenShift
Origin with the EFK stack from origin-aggregated-logging? We are
looking for an upstream home for that testing. We are currently using
ci.openshift.redhat.com for logging testing:
https://ci.openshift.redhat.com/jenkins/job/test-origin-aggregated-logging/249/
There is also potentially overlap with the ConfigManagement SIG
here as well.
It is intended that this will form the basis of
https://github.com/openshift/origin-aggregated-logging which
will be built from the packages and base images provided by
the SIG.
If you are interested, please chime in in the email thread:
https://lists.centos.org/pipermail/centos-devel/2016-May/014777.html
--
Jason DeTiberus
--
Jason DeTiberus
_______________________________________________
users mailing list
[email protected]
http://lists.openshift.redhat.com/openshiftmm/listinfo/users
