GeoServer will happily continue returning exception documents when it 
has lost its database connection, for example. Tomcat will be happy, the 
user not so happy.

Remote monitoring will detect a fully hung server. A monitoring script 
on the server will also hang if the server is hung, so you'll never know 
about it. If the provider is having a network outage, you will not 
receive any notifications if your monitoring is on the same network.

Nagios also provides features such as flap detection as well as 
plotting, which is useful to spot trends and look for past outages. It 
is also widely-used and likely better tested than a custom script.

I have not used Nagios for remote restart (what scares me is the risk of 
Nagios misconfiguration bringing down a production system), but I expect 
that it could be used for it. Might even be able to use it to restart a 
virtual machine in a cluster. It is pretty extensible.

In short, what are the availability expectations? What types of failures 
do you wish to guard against?

On 27/03/12 16:57, Frank Gasdorf wrote:
> I agree, we use Nagios in our infrastructure as well, but in this context I 
> seems to me that the server is hosted by an external provider and I guess its 
> a bit overkill to install Nagios only for the simple propose to watching and 
> alerting. The script itself has the advantage to restart the server, which 
> Nagios would not have ;)

-- 
Ben Caradoc-Davies <[email protected]>
Software Engineer
CSIRO Earth Science and Resource Engineering
Australian Resources Research Centre

------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here 
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
Geoserver-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geoserver-users

Reply via email to