Fair Proxy Balancer

[David Pratt]

Hi Rt. There are broader implications than ec2. There is already 
competition for on demand commodity computing and this will only 
increase. It really is not important if it is ec2, a competitive 
service, or virtual servers on a dedicated host, or your own equipment. 
Monitoring, assessing and responding to changing circumstances is all 
part maintaining a deployment.

For ec2 in particular, the popular solution to a poor performing 
instance is to terminate the instance and start another. It is just as 
important to know when you may wish to scale or drop instances.

The potential api that has been discussed will be welcome. Direct 
measurement from NGINX will be better that reporting results that have 
occurred some time in the past.

Regards,
David

Rt Ibmer wrote:
> David - so you are using ec2 as well? If you go back over the past couple of weeks you will see our discussions here on enhancing nginx's monitoring capabilities.
> 
> For instance one thing I would like is to have nginx somehow notify an external program in the event it deems an upstream to be offline.  This way such a script could apply other logic and decide to launch a replacement upstream automatically.
> 
> The latest school of thought we've had on this is that nginx may writes status information into a shared memory location where a monitoring script could keep an eye out and get reports.
> 
> Francois - it would be great if one of the things nginx could report on would be things like average connection and response times on a per upstream basis.  Is this doable?
> 
> 
>       ____________________________________________________________________________________
> Be a better friend, newshound, and 
> know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
> 
>