kefoster
Contributor
Contributor

Scalability of HQ

Howdy,

I am a sys admin in a development environment of 900+ servers. We
currently use Nagios to monitor our systems in use which is about
830. I was wondering what the scalability of HQ is?

Our current HQ server:
Dell 1750, Dual 3Ghz, 4GB mem
OS: Red Hat 4.0
The HQ install is running on the internal database server but if we
decide to go with HQ we will move it to Oracle.

Any information on this would be appreciated.

Also, good job on the product! We like what we have seen so far.

Thank you,

Kevin Foster

0 Kudos
3 Replies
admin
Immortal
Immortal

Hi Kevin,

Thank you for your interest in Hyperic HQ. The scalability of HQ
depends on numerous factors.

- Number of hardware platforms being monitored
- Number of resources on each platform being monitored
- Number and collection intervals of metrics
- Number of alert definitions
- Database performance and throughput

Each of these factors, of course, can be determined by the user. Our
internal database is actually PostgreSQL, which is not a limiting
factor. We have customers running large deployments for datacenters
of hundreds of platforms on both PostgreSQL and Oracle. I am sure
that HQ can scale to fit your environment. However, we would need to
determine your specific monitoring and management requirements and
see if any tradeoffs need to be made to gain the coverage that you
need. I will say that for those customers with larger deployments,
the database is typically running on a separate dedicated server and
that the HQ server has typically around 16GB of memory.

We also have the ability to create a high-availability environment
for our enterprise customers by creating a farm of data processing
nodes that partition your server farm to report to different data
processors. This can significantly improve HQ's capacity. This
will, of course, involve a few more steps in the installation and
configuration of HQ. I hope this helps to answer your question.
Feel free to ping me with any further questions. Thanks!

Charles


On Jun 9, 2006, at 12:48 PM, Kevin Foster wrote:

> Howdy,
>
> I am a sys admin in a development environment of 900+ servers. We
> currently use Nagios to monitor our systems in use which is about
> 830. I was wondering what the scalability of HQ is?
>
> Our current HQ server:
> Dell 1750, Dual 3Ghz, 4GB mem
> OS: Red Hat 4.0
> The HQ install is running on the internal database server but if we
> decide to go with HQ we will move it to Oracle.
>
> Any information on this would be appreciated.
>
> Also, good job on the product! We like what we have seen so far.
>
> Thank you,
>
> Kevin Foster


0 Kudos
kefoster
Contributor
Contributor


Thank you for the information!  We really only plan to monitor availability, CPU, Memory, Drive Space, and Networking for the majority of the systems.  We have an infrastructure that supports the development so those would be a little more involved.

The auto discovery is going to find a lot of the App Servers in our environment.  We write apps that run on top of, bea, websphere, sun app server, tomcat... you get the idea.  Is there a way to turn off auto discovery for certain pieces?

Thanks,

Kevin Foster

On 6/9/06, Charles Lee <clee@hyperic.com> wrote:
Hi Kevin,

Thank you for your interest in Hyperic HQ.  The scalability of HQ
depends on numerous factors.

- Number of hardware platforms being monitored
- Number of resources on each platform being monitored
- Number and collection intervals of metrics
- Number of alert definitions
- Database performance and throughput

Each of these factors, of course, can be determined by the user.  Our
internal database is actually PostgreSQL, which is not a limiting
factor.  We have customers running large deployments for datacenters
of hundreds of platforms on both PostgreSQL and Oracle.  I am sure
that HQ can scale to fit your environment.  However, we would need to
determine your specific monitoring and management requirements and
see if any tradeoffs need to be made to gain the coverage that you
need.  I will say that for those customers with larger deployments,
the database is typically running on a separate dedicated server and
that the HQ server has typically around 16GB of memory.

We also have the ability to create a high-availability environment
for our enterprise customers by creating a farm of data processing
nodes that partition your server farm to report to different data
processors.  This can significantly improve HQ's capacity.  This
will, of course, involve a few more steps in the installation and
configuration of HQ.  I hope this helps to answer your question.
Feel free to ping me with any further questions.  Thanks!

Charles


On Jun 9, 2006, at 12:48 PM, Kevin Foster wrote:

> Howdy,
>
> I am a sys admin in a development environment of 900+ servers.  We
> currently use Nagios to monitor our systems in use which is about
> 830.  I was wondering what the scalability of HQ is?
>
> Our current HQ server:
> Dell 1750, Dual 3Ghz, 4GB mem
> OS: Red Hat 4.0
> The HQ install is running on the internal database server but if we
> decide to go with HQ we will move it to Oracle.
>
> Any information on this would be appreciated.
>
> Also, good job on the product!  We like what we have seen so far.
>
> Thank you,
>
> Kevin Foster



0 Kudos
admin
Immortal
Immortal


Hi Kevin,

This is easy to do.  There are several ways of doing this.

1) Prior to starting your agent, add in a directive in agent.properties file for plugins.exclude.  So for example, to exclude Weblogic, Websphere, and Tomcat, you would have the following in agent.properties:

plugins.exclude=weblogic,websphere,servlet

Notice that the names are the jar filenames in [agent install dir]/pdk/lib minus the -plugin.jar part.

2) Prior to starting your agent, remove the plugin jar files out of [agent install dir]/pdk/lib that you do not want to auto-discover.  Of course, if you were to change your mind, it's more of a hassle to have to replace those files.

3) You can choose not to import them.  From the dashboard, you can check the discovered servers and click on the REMOVE button to never see them again.  However, you can only do this for one platform at a time.

I hope this helps.

Charles



On Jun 9, 2006, at 7:03 PM, Kevin Foster wrote:

Thank you for the information!  We really only plan to monitor availability, CPU, Memory, Drive Space, and Networking for the majority of the systems.  We have an infrastructure that supports the development so those would be a little more involved.

The auto discovery is going to find a lot of the App Servers in our environment.  We write apps that run on top of, bea, websphere, sun app server, tomcat... you get the idea.  Is there a way to turn off auto discovery for certain pieces?

Thanks,

Kevin Foster

On 6/9/06, Charles Lee <clee@hyperic.com> wrote:
Hi Kevin,

Thank you for your interest in Hyperic HQ.  The scalability of HQ
depends on numerous factors.

- Number of hardware platforms being monitored
- Number of resources on each platform being monitored
- Number and collection intervals of metrics
- Number of alert definitions
- Database performance and throughput

Each of these factors, of course, can be determined by the user.  Our
internal database is actually PostgreSQL, which is not a limiting
factor.  We have customers running large deployments for datacenters
of hundreds of platforms on both PostgreSQL and Oracle.  I am sure
that HQ can scale to fit your environment.  However, we would need to
determine your specific monitoring and management requirements and
see if any tradeoffs need to be made to gain the coverage that you
need.  I will say that for those customers with larger deployments,
the database is typically running on a separate dedicated server and
that the HQ server has typically around 16GB of memory.

We also have the ability to create a high-availability environment
for our enterprise customers by creating a farm of data processing
nodes that partition your server farm to report to different data
processors.  This can significantly improve HQ's capacity.  This
will, of course, involve a few more steps in the installation and
configuration of HQ.  I hope this helps to answer your question.
Feel free to ping me with any further questions.  Thanks!

Charles


On Jun 9, 2006, at 12:48 PM, Kevin Foster wrote:

> Howdy,
>
> I am a sys admin in a development environment of 900+ servers.  We
> currently use Nagios to monitor our systems in use which is about
> 830.  I was wondering what the scalability of HQ is?
>
> Our current HQ server:
> Dell 1750, Dual 3Ghz, 4GB mem
> OS: Red Hat 4.0
> The HQ install is running on the internal database server but if we
> decide to go with HQ we will move it to Oracle.
>
> Any information on this would be appreciated.
>
> Also, good job on the product!  We like what we have seen so far.
>
> Thank you,
>
> Kevin Foster




0 Kudos