[ 
https://issues.apache.org/jira/browse/HDDS-3006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17039057#comment-17039057
 ] 

Marton Elek commented on HDDS-3006:
-----------------------------------

Not sure what is this issue about, I checked the the design doc but didn't get 
the idea from there.

I don't think that we need to take care of installing any components, including 
prometheus. It seems to be the responsibility of other, external provisioners: 
for such as Kubernetes, Cloudera Manager or Ambari. I couldn't imagine how can 
we cover all the use cases with provide some very basic install process 
(especially as we need to support containerized and non-containerized 
environments).

I propose to use any existing Prometheus as is. There can be a configuration 
option which defines the location (and credentials?) of the Prometheus instance 
which can be used. We already have Prometheus in the docker-compose definition 
which is ready to use, we can adjust the configuration to always have it when 
we have recon.



> Integrate Prometheus into Ozone Recon.
> --------------------------------------
>
>                 Key: HDDS-3006
>                 URL: https://issues.apache.org/jira/browse/HDDS-3006
>             Project: Hadoop Distributed Data Store
>          Issue Type: Sub-task
>          Components: Ozone Recon
>    Affects Versions: 0.5.0
>            Reporter: Aravindan Vijayan
>            Assignee: Attila Doroszlai
>            Priority: Major
>             Fix For: 0.5.0
>
>
> The goal of Recon is to provide a window into the state of the Ozone cluster. 
> Given that it will have OM and SCM metadata,  the next step is to get 
> monitoring information from the ozone components. This will help Recon obtain 
> more dynamic information about the active pipelines and failed transactions. 
> Since all the ozone components are already compatible with pushing metrics to 
> a prometheus end point, we can go ahead with that as the metric store. 
> The tasks involved in this JIRA are :
> * Recon deploy should bring in a prometheus instance with it.
> * Recon start/stop will manage the prometheus start/stop.
> Since this instance is only expected to be used by Ozone, we don't need to 
> have it fully customizable. A default prometheus config file can be included 
> that can be used by the prometheus instance.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to