Open Stack

Hello, Girija!
Read some information about Ceilometer here: [1].
In short, it works in two ways:
1. Collecting information with the help of several "pollsters" doing
Openstack API calls.
2. Listening Openstack notification bus for incoming messages.
In both ways you need something that will gather information and put it
straight to the pipeline (1st way), or form a message and send it to the
messaging bus (2nd way).
I think it might be possible to gather information you need with some
monitoring tool like Zabbix and then, using second approach, send it to
messaging system. To do that you'll need to:
1. Write a plugin for that monitoring tool that will push messages.
2. Write plugin for Ceilometer, which will listen for that messages, gather
them and put into pipeline.
3. Configure pipeline to accept those messages and publish them (to DB,
file, etc).
Please take a look at conversation [2] which probably will make you give up
the idea of using Ceilometer for anything that should be "live". It's easy
to post tons of information to Ceiloemter, but hard to get them back
quickly. Also take a look at Gnocchi project [3].
Hope that helps a bit.

[1]
http://docs.openstack.org/developer/ceilometer/architecture.html#gathering-the-data
[2] http://www.gossamer-threads.com/lists/openstack/operators/44584
[3] http://docs.openstack.org/developer/gnocchi/


On Fri, Jun 5, 2015 at 7:50 AM Girija Sharan <girijasharansingh at gmail.com>
wrote:

> Hi All,
>
> Please do reply guys.
>
> On Thu, Jun 4, 2015 at 11:38 AM, Girija Sharan <
> girijasharansingh at gmail.com> wrote:
>
>> Hi All,
>>
>> This query is about Ceilometer, and its my sincere request to you to read
>> this till end and understand properly.
>>
>> Let me first put up my requirement :
>> 1). I need to develop a monitoring tool for my infrastructure, which will
>> show me the live CPU usage, Memory usage, Disk usage, and Network traffic
>> of instances launched on Compute nodes as well as of physical Compute nodes.
>>
>> To achieve this I explored Ceilometer service of Openstack (metering part
>> only) and found that, it has many meters using which I can retrieve above
>> mentioned live data but only of instances launched, not of physical Compute
>> nodes.
>>
>> Now moving to my queries :
>> 1). How do I create corresponding meters to retrieve above mentioned
>> details of physical Compute nodes (like memory.usage, cpu_util,
>> network.incoming.bytes, etc.. which exists for instances).
>> 2). Please do point me towards some helpful links or let me know the
>> exact way of doing so.
>> 3). Why don't we add Monitoring service to Openstack. As Admin user will
>> be needing to see all these details related to Compute nodes and other
>> nodes.
>>
>> Now been said that, I came up with an idea which I would like to share
>> here :
>>
>> As part of VM migration, one has to know the load on Compute nodes,
>> depending on which Admin will be deciding whether to migrate the VMs or not
>> and to which Compute node.
>> Now just think if VM migration need to be done automatically (when some
>> specified threshold value reaches), one need to collect these metrics from
>> Compute nodes.
>>
>> We can make use of these existing Ceilometer's meters (metering part of
>> Telemetry) and adding more additional meters to retrieve details from
>> Physical nodes, to create Monitoring Service as well.
>>
>> Thanks in advance and early responses will be highly appreciated.
>>
>>
>> Thanks and Regards,
>> Girija Sharan Singh
>>
>
> _______________________________________________
> Mailing list:
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
> Post to     : openstack at lists.openstack.org
> Unsubscribe :
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20150605/56fe1ada/attachment.html>