cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sebastian Gomez <tioc...@gmail.com>
Subject Re: Cloud Utilization
Date Thu, 22 Jun 2017 21:04:43 GMT
First of all, what means "instance utilization is 0%", just kernel
processes and maintenance tasks can increase CPU usage during this period.
No utilization means no logins?

I don't know any service/soft that do it, but you can do it yourself:

For example:
You can easily develop a script to be executed daily on each server to
detect if there is or not usage (whatever it means), and register into a
file (p.e. a counter). One the counter reaches the maximum (p.e. 3 days),
shutdown itself.
For other side, develop a script in other server to make queries on the
cloud database or cloud usage database to ask for instances which have more
than maximum days without being powered on, and call the API method to
destroy them.

Thinking about it I'm sure that people will find better ways to do it.
If you build the solution, please, consider sharing it!


Good luck.






Atentamente,
Sebastián Gómez

On Wed, Jun 14, 2017 at 9:15 AM, Ivan Kudryavtsev <kudryavtsev_ia@bw-sw.com>
wrote:

> Hello, Ghaith.
>
> I think ACS doesn't support this easily, but It still can be achieved. The
> second part of your demand is quite simple and probably can be achieved
> with just API VM list call and additional information from the database,
> but the first one is out of ACS features obviosly.
>
> I just would like to direct you to a half-step solution which can be
> extended to achieve desired outcome.
> I don't know what hypervisor you use, but our the solution is for KVM.
>
> Take a look at https://github.com/bwsw/cs-pulse-sensor - this is libvirt
> sensor which stores metrics for all VMs to the InfluxDB database which
> later can be asked for time-series like "show me VMs with low CPU
> utilization for a period of time" and further stop that VM with API. We
> also have https://github.com/bwsw/cs-pulse-server/tree/develop which shows
> how to request the database and get CPU utilization results.
>
> Unfortunately, I don't use other hypervisors and they probably require
> other sensors than libvirt.
>
> So, finally, it's not so difficult to implement the feature you need but
> requires a custom development.
> Still, if you need that custom development, please contact me and we can
> discuss.
>
>
> 2017-06-14 13:28 GMT+07:00 Ghaith Bannoura <gbannoura@etq.com>:
>
> > Hello All,
> >
> > I have a case that I need to turn off instance in case that the instance
> > utilization is 0%  for three days ( the instance is not used ) and
> destroy
> > the instance if its stopped more than 7 days .
> >
> > Is there any automated way to do like this case (like Ansible) , I
> believe
> > that auto scaling while handle turn off instance , but I need another
> > solution
> >
> > Best Regards,
> > Ghaith Bannoura
> > Senior System Administrator
> > MCT, MCSE (Messaging, Server Infrastructure)
> > MCSA (Windows Server 2008, 2012), MCP
> >
> >
> >
>
>
> --
> With best regards, Ivan Kudryavtsev
> Bitworks Software, Ltd.
> Cell: +7-923-414-1515
> WWW: http://bitworks.software/ <http://bw-sw.com/>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message