Hi Ethan,
I agree that the service should be separated from each other,
and requesting a new repo is also a valid option. Thanks for your suggestion.
Best Regards,
Bin Yang, Solution Engineering Team, Wind River
ONAP Multi-VIM/Cloud PTL
Direct +86,10,84777126 Mobile +86,13811391682 Fax +86,10,64398189
Skype: yangbincs993
From: [email protected] [mailto:[email protected]] On
Behalf Of ethanlynnl
Sent: Thursday, September 06, 2018 1:55 PM
To: onap-discuss; Yang, Bin
Cc: Addepalli, Srinivasa R; [email protected]; Multanen, Eric
W; Ranganathan, Dileep; Ramki Krishnan; Ranganathan, Raghu
Subject: Re: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Bin,
I think it’s better to use a new repository to host HPA telemetry, since the
broker are written in python and we need to consider the performance. If
massing broker with telemetry, massive telemetries might downgrade the broker’s
performance.
On 6 Sep 2018, at 9:33 AM, Yang Bin
<[email protected]<mailto:[email protected]>> wrote:
Hi Srini,
I double checked that during last MultiCloud weekly meeting, and
I didn’t hear any objection. So I am thinking it is okay to use multicloud
framework repository to host the HPA telemetry.
However, before anyone starts the patch upstreaming, I would like to know the
details of your seed codes, and see how it fit into the existing framework repo
since there is already source code for multicloud broker there. We want to keep
thing consistent and easy /straight forward to understand/maintain.
So I suggest that you or someone else will showcase the seed code to multicloud
team and then decide how it should be proceeded.
Thanks.
Best Regards,
Bin Yang, Solution Engineering Team, Wind River
ONAP Multi-VIM/Cloud PTL
Direct +86,10,84777126 Mobile +86,13811391682 Fax +86,10,64398189
Skype: yangbincs993
From: Addepalli, Srinivasa R [mailto:[email protected]]
Sent: Thursday, September 06, 2018 1:26 AM
To: [email protected]<mailto:[email protected]>; Yang, Bin;
[email protected]<mailto:[email protected]>
Cc: Multanen, Eric W; Ranganathan, Dileep;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
Subject: RE: [onap-discuss] [MULTICLOUD] - HPA telemetry
Thanks Bin. Let us know if everybody is okay to use framework repository or
provide suggestions.
HPA telemetry is part of HPA umbrella and Alex presented to various committees
for R3 additions. Dileep is working on A&AI schema changes and related changes
needed in OOF. But in these architecture/use-case meetings, we don’t get into
the details of repositories in each project and hence that part was not
discussed. In both A&AI and OOF, no need to create any repositories. I hope
that we don’t need to create any repository here too. My view is that
framework repo seems to be generic and since HPA telemetry is also generic,
related software can be placed there. Let us know.
Thanks
Srini
From: [email protected]<mailto:[email protected]>
[mailto:[email protected]] On Behalf Of
[email protected]<mailto:[email protected]>
Sent: Tuesday, September 4, 2018 8:18 PM
To: [email protected]<mailto:[email protected]>; Addepalli,
Srinivasa R
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>
Cc: Multanen, Eric W
<[email protected]<mailto:[email protected]>>; Ranganathan,
Dileep <[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
<[email protected]<mailto:[email protected]>>
Subject: Re: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Srini,
Thanks for your feedback. I will check with our team to see if
any further question/concerns with regarding to this proposal.
It seems it is a new feature impacting multiple ONAP projects (MultiCloud, AAI,
OOF, more? ), so I would also like to know whether this proposal has been
presented to ARC subcommittee? What is the suggestion there?
Thanks
Best Regards,
Bin Yang, Solution Engineering Team, Wind River
ONAP Multi-VIM/Cloud PTL
Direct +86,10,84777126 Mobile +86,13811391682 Fax +86,10,64398189
Skype: yangbincs993
From: [email protected]<mailto:[email protected]>
[mailto:[email protected]] On Behalf Of Srini
Sent: Wednesday, September 05, 2018 3:24 AM
To: Yang, Bin; [email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>
Cc: Multanen, Eric W; Ranganathan, Dileep;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
Subject: Re: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Bin,
I try to answer some of the questions inline.
In regards to repository, I see following:
<image001.png>
Since stats aggregation service is common across Cloud technologies, I think
framework repository is a good candidate for adding this generic code and
rules. What do you think?
From: [email protected]<mailto:[email protected]>
[mailto:[email protected]]
Sent: Thursday, August 30, 2018 7:09 PM
To: Addepalli, Srinivasa R
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>
Cc: Multanen, Eric W
<[email protected]<mailto:[email protected]>>; Ranganathan,
Dileep <[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
<[email protected]<mailto:[email protected]>>
Subject: RE: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Srini,
We had conduct an Q&A session during last MultiCloud weekly
meeting , the purpose is to dive into more details of intent/design of this HPA
telemetry so that we can judge whether it fits into multicloud’s scope and how
it should be incubated inside Multicloud . Please refer to the MOM
(https://wiki.onap.org/display/DW/MOM+of+Aug.+29th+2018+MultiCloud+weekly+meeting<https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwiki.onap.org%2Fdisplay%2FDW%2FMOM%2Bof%2BAug.%2B29th%2B2018%2BMultiCloud%2Bweekly%2Bmeeting&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974449680&sdata=sv3nDZm%2BYMpeKbF4By0dXbc7HtpmlInJdk8VLoSCfwc%3D&reserved=0>
) for details, I also copied the questions/answers here to facilitate the
further discussion/feedback:
HPA Telemetry
Q&A:
Q: Delivery plan:
A: Start it in C. Release, no commitment what can be delivered
SRINI> True as we have started this work late. But, minimum we want to do in
Casablanca is to integrate metrics collection, provide rules to aggregate the
information and provide visualization via Grafana. Our stretch goal is to
populate HPA resource information in A&AI.
Q: Seed code:
A: N/A
Q: Deployment topology: will be a single collector running for any kind of
underlying VIM/Cloud type.instance?
A: Not clear yet.
SRINI> Yes. It is Cloud technology agnostic. As far as service provider has a
way to install exporters in compute nodes, this should work irrespective of
whether the cloud technology is VIO, Titanium, upsteam openstack orK8S.
Q: Configuration API/portal:
A: N/A for now, but eventually be there,
Q: Will the API/IM be generic to be VIM/Cloud agnostic ?
A: not clear yet
SRINI> Yes. Please see above comment.
Q: Is there any dependency to AAI model/schema
A: There is a schema change to AAI in progress (Dileep)
Q: Is this schema be generic (VIM/Cloudagnostic?)
A: Supposed to be generic
SRINI> Yes. It is generic.
Q: Will the collector? Be impacted by different agent on VIM/Cloud ?
A: not clear yet
SRINI> Current support is for Prometheus exporters as well as CollectD. But,
it should not be a matter as long as the agent supports exporting metrics as
expected by Prometheus service.
Q: Why not DCAE VES collector /microservice? Please evaluate this option.
A: no answer yet
SRINI> This will be below DCAE. Intention (as roadmap) to send
alerts/events/aggregation-data to DCAE via VES.
suggestion (Bin
Yang<https://na01.safelinks.protection.outlook.com/?url=https:%2F%2Fwiki.onap.org%2Fdisplay%2F~biny993&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974449680&sdata=cYy3zZLumr04raKcwPLly17CJ9g1s%2FeLZJwavq6unEA%3D&reserved=0>
) If there can be just 1 collector for all VIM/Cloud instance/types, you can
have dedicate repo
Otherwise, share the existing repos following the broker/plugin topology
AI: Eric
Multanen<https://na01.safelinks.protection.outlook.com/?url=https:%2F%2Fwiki.onap.org%2Fdisplay%2F~ewmulta&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974459685&sdata=DdqExjdHhmb9yjJi4vUWRssiQfCVrZrz3v6tEQ8PxwU%3D&reserved=0>
figure out the questions above, then decide whether a dedicated repo needed.
Best Regards,
Bin Yang, Solution Engineering Team, Wind River
ONAP Multi-VIM/Cloud PTL
Direct +86,10,84777126 Mobile +86,13811391682 Fax +86,10,64398189
Skype: yangbincs993
From: Addepalli, Srinivasa R [mailto:[email protected]]
Sent: Friday, August 31, 2018 12:51 AM
To: [email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>;
Yang, Bin
Cc: Multanen, Eric W; Ranganathan, Dileep;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
Subject: RE: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Frank,
Please see following links:
Intent and high level architecture is presented in various groups:
https://wiki.onap.org/display/DW/Edge+Scoping+MVP+for+Casablanca+-+ONAP+Enhancements#EdgeScopingMVPforCasablanca-ONAPEnhancements-AggregatedInfrastructureTelemetryStreams(AlignswithHPArequirements,CombiningeffortswithHPA)<https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwiki.onap.org%2Fdisplay%2FDW%2FEdge%2BScoping%2BMVP%2Bfor%2BCasablanca%2B-%2BONAP%2BEnhancements%23EdgeScopingMVPforCasablanca-ONAPEnhancements-AggregatedInfrastructureTelemetryStreams(AlignswithHPArequirements%2CCombiningeffortswithHPA)&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974459685&sdata=aRyw2k1b0tw9JJURrXCCl1ocRkdQwyjPjsZs8ZCh9Go%3D&reserved=0>
Dileep started design page at (Telemetry for OOF and A&AI)
https://wiki.onap.org/pages/viewpage.action?pageId=38116734<https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwiki.onap.org%2Fpages%2Fviewpage.action%3FpageId%3D38116734&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974469694&sdata=1hOzBft%2BScZj4fRRBmc1j1u03nOtqWBBloZt1poQWww%3D&reserved=0>
There is no design page yet for HPA Telemetry for Multi-Cloud. But there are
set of JIRA request on the following EPIC:
https://jira.onap.org/browse/MULTICLOUD-257<https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fjira.onap.org%2Fbrowse%2FMULTICLOUD-257&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974469694&sdata=HbiKhVJ5E%2FPDGY8C6SaxFuSZpRexAJ4u18%2Fe5Qu9KJs%3D&reserved=0>
Thanks
Srini
From: [email protected]<mailto:[email protected]>
[mailto:[email protected]] On Behalf Of Frank Sandoval
Sent: Wednesday, August 29, 2018 12:51 PM
To: onap-discuss
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>
Cc: Addepalli, Srinivasa R
<[email protected]<mailto:[email protected]>>;
Multanen, Eric W <[email protected]<mailto:[email protected]>>;
Ranganathan, Dileep
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
<[email protected]<mailto:[email protected]>>
Subject: Re: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Srini,
This topic has been discussed in the OOF group as well. Is there a wiki page,
slide deck, or other material describing the proposed design? Thanks
Frank Sandoval
[email protected]<mailto:[email protected]>
OAM Technologies, representing Arm
OOF committer
On Aug 28, 2018, at 11:51 PM, Yang Bin
<[email protected]<mailto:[email protected]>> wrote:
Hi Srini,
If possible, let’s continue the discussion on the upcoming
MultiCloud weekly meeting.
Thanks.
Best Regards,
Bin Yang, Solution Engineering Team, Wind River
ONAP Multi-VIM/Cloud PTL
Direct +86,10,84777126 Mobile +86,13811391682 Fax +86,10,64398189
Skype: yangbincs993
From: Addepalli, Srinivasa R [mailto:[email protected]]
Sent: Friday, August 24, 2018 12:32 AM
To: Multanen, Eric W;
[email protected]<mailto:[email protected]>; Yang, Bin
Cc: Ranganathan, Dileep; [email protected]<mailto:[email protected]>;
Ranganathan, Raghu
Subject: RE: [onap-discuss] [MULTICLOUD] - HPA telemetry
Thanks Eric.
Few more answers embedded below.
From: Multanen, Eric W
Sent: Wednesday, August 22, 2018 5:01 PM
To: Addepalli, Srinivasa R
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>
Cc: Ranganathan, Dileep
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
<[email protected]<mailto:[email protected]>>
Subject: RE: [onap-discuss] [MULTICLOUD] - HPA telemetry
Bin,
Thank you for the agenda slot today, I think we covered most of everything –
the call dropped promptly on the hour.
Following are the questions I collected from the discussion – please amend and
clarify if I haven’t
captured your key questions.
I expect Srini and others can assist in providing more detailed/correct answers.
1. Why is this part (Prometheus) of ONAP/multi-cloud instead of Openstack?
a. it is used more broadly than just for Openstack
SRINI>
- This is meant for infrastructure and HPA metrics.
- We want to be agnostic to VIM technology
- Many constrained Edge deployments would like minimal functions in the
Edge – Like nova, neutron, Cinder, HEAT and leave rest to be put in elsewhere.
- Prometheus is second only to K8S that got graduated in CNCF (only two
so far).
- Prometheus has very good integration with Openstack & K8S, AWS, AZURE,
Bare-metal nodes including collected.
2. What is the relationship between Openstack telemetry and this proposal?
SRINI> Prometheus is mainly meant to collect the metric, aggregate/summation,
siliencing etc… Prometheus even have a way collect the metric from Openstack.
3. What is the interface and/or API for configuring rules ?
a. Initially, configuration files for the service.
4. How are different cloud / infrastructure type supported?
SRINI> Fortunately, Prometheus has various integration with popular cloud
technologies already.
5. Where is the datastore of the Prometheus service ? What data/info is
accessible by ONAP?
a. plan is to develop the HPA statistics exporting/receiving service to
provide specific HPA data to ONAP.
b. not sure yet if there is a generic access to all data (and whether that
is desired)
Eric
From: Addepalli, Srinivasa R
Sent: Tuesday, August 21, 2018 9:36 AM
To: [email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>
Cc: Multanen, Eric W
<[email protected]<mailto:[email protected]>>; Ranganathan,
Dileep <[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>; Ranganathan, Raghu
<[email protected]<mailto:[email protected]>>
Subject: RE: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Bin,
I am travelling. Eric confirmed and he will talk about the feature and high
level design. If time permits, Dileep can show the A&AI schema proposal.
Hi Ramki and Raghu,
Since it is born out of Edge-automation working group, it would be good if you
can attend the meeting.
Thanks
Srini
From: [email protected]<mailto:[email protected]>
[mailto:[email protected]] On Behalf Of
[email protected]<mailto:[email protected]>
Sent: Tuesday, August 21, 2018 8:38 AM
To: [email protected]<mailto:[email protected]>; Addepalli,
Srinivasa R
<[email protected]<mailto:[email protected]>>
Cc: Multanen, Eric W
<[email protected]<mailto:[email protected]>>; Ranganathan,
Dileep <[email protected]<mailto:[email protected]>>
Subject: Re: [onap-discuss] [MULTICLOUD] - HPA telemetry
Hi Srini,
Before I can suggest, I would like to understand this feature more.
Would someone please present the idea/design to MultiCloud team? Perhaps it's
a good chance to do that during the upcoming MultiCloud weekly meeting.
Thanks
Bin
在 2018年8月21日，23:29，Srini
<[email protected]<mailto:[email protected]>> 写道：
Hi Bin,
We have this EPIC as part of Edge automation:
https://jira.onap.org/browse/MULTICLOUD-257<https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fjira.onap.org%2Fbrowse%2FMULTICLOUD-257&data=02%7C01%7Cethanlynnl%40vmware.com%7Cf17f054c8e2d42b2af7608d61398b687%7Cb39138ca3cee4b4aa4d6cd83d9dd62f0%7C1%7C0%7C636717943974479699&sdata=4LsY0CqOJvkO9V1jEx7MRC7MC0oaPWSOHWbbrqR6nSY%3D&reserved=0>
This is meant to ensure that placement decisions consider current available
resources in addition to what we were doing with respect to capabilities till
R2.
This work got started last week.
There would some source code development (We call it – HPA Telemetry receiving
service) which gets aggregated telemetry information, do any
massaging/filtering necessary and updates A&AI DB.
There would be some additional development with respect rules for Prometheus
aggregation service.
This is generic service required across Openstack, K8S and Azure. Is there any
repository in Multi-Cloud we can use to put this code or do you suggest to
request for a new repository? Please advise.
Thanks
Srini
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#12260): https://lists.onap.org/g/onap-discuss/message/12260
Mute This Topic: https://lists.onap.org/mt/24877170/21656
Group Owner: [email protected]
Unsubscribe: https://lists.onap.org/g/onap-discuss/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
