[OASIS Issue Tracker] (TOSCA-216) General Issues for Monitoring Ad Hoc Group

[OASIS Issues Tracker <workgroup_mailer@lists.oasis-open.org>]

    [ https://issues.oasis-open.org/browse/TOSCA-216?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=55949#comment-55949 ] 

Chip Holden commented on TOSCA-216:
-----------------------------------

Answer from Ifat:

1, in the proposal, does one  metric sample only monitoring one parameter for the application?
If I would like to monitor for example, cpu, memory as well as some application KPIs, do I need to establish many sample metrics, and one for each parameters?
=========
[Ifat]: Yes, I think that for every parameter there should be a different metric sample. You might want to do additional aggregations based on the metric samples, like average or aggregation over time, and it doesn’t make sense to mix different types of metrics (like CPU and MEMORY) in these aggregations.
=========
 
2, how do I know which parameter (or parameters in case for my first commends that one metric sample can monitoring several parameters) the sample metric monitoring for? In the tosca.monitoring.MetricSample definition in A.1.3.1, only in the description element say that “A single metric sample,applicatio KPI, like CPU, MEMORY, etc”, but there is no other place in the sample saying what exactly this sample metric monitoring for, cpu or memory? Maybe in the monitoring_endpoint capability, it needs to indicate what parameter does this endpoint expose to the outside for monitoring, so when a sample metric connect to this endpoint, it means that this sample metric is for that paramter.
=========
[Ifat]: Indeed the monitoring endpoint should hold the exact definition of which parameter is being monitored, and how to get it. The monitoring endpoint is not fully defined yet, we only have a general idea of what it should look like. For example, in Alcatel-Lucent we defined an endpoint that executes a script that returns the metric value, but there may be other types of endpoints (HTTP, SQL, etc.).
In addition, there are different types of metrics:
1.       Application-specific metric, which are defined by the monitoring endpoints

2.       Resource consumption metric, like CPU, MEMORY, etc. We need to decide how to define these metrics



> General Issues for Monitoring Ad Hoc Group 
> -------------------------------------------
>
>                 Key: TOSCA-216
>                 URL: https://issues.oasis-open.org/browse/TOSCA-216
>             Project: OASIS Topology and Orchestration Specification for Cloud Applications (TOSCA) TC
>          Issue Type: Improvement
>          Components: Monitoring
>            Reporter: Paul Lipton 
>            Assignee: Chip Holden
>




--
This message was sent by Atlassian JIRA
(v6.2.2#6258)