Nagios Support Forum

Hello NAGIOS XI team

Please advice what is the best way to monitor REDHAT 6.x enterprise clusters and get their performance/capacity reported "as one", not as break-down per node?

Thank you

Have you searched our Exchange, for a plugin that does this? There seem to be a couple towards the bottom of this page, https://exchange.nagios.org/index.php?o ... rd=cluster

Are any of these what you're looking for?

rkennedy wrote:Have you searched our Exchange, for a plugin that does this? There seem to be a couple towards the bottom of this page, https://exchange.nagios.org/index.php?o ... rd=cluster

Are any of these what you're looking for?

I am more and more worried that our move towards NAGIOX XI was a mistake to begin with:
- your general approach to clusters is a simple nodes health plus services running, but not components; while we need something that Zabbix does: representation of cluster as a single entity and thus proper capacity/performance reporting.

Is there anything like this Nagios XI could offer? (really same goes for Windows clusters too)

dlukinski wrote:your general approach to clusters is a simple nodes health plus services running, but not components

The issue with this statement is that, strictly speaking, Nagios XI doesn't really *have* an approach - we can monitor anything as long as there is a plugin for it. If there is a plugin that considers the cluster as a whole then we can absolutely do that. There may well be a plugin on our Exchange that does in fact monitor the cluster as you describe. Otherwise...

dlukinski wrote:we need [...] representation of cluster as a single entity and thus proper capacity/performance reporting.

If a plugin does not currently exist for this, you can use something like BPI (Business Process Intelligence) to logically group hosts/services together into a "meta-group" so to speak, such that if X number of the members are in a non-OK state you can alert based off of that instead of any individual member's status.

https://assets.nagios.com/downloads/nag ... _Addon.pdf

Over the years I have found that Nagios XI monitors about 90% of what people typically need out of the box, another 9% can be monitored using plugins available on our Exchange (or elsewhere), and for about 1% of things there is not a plugin readily-available. I don't know of any software out there that can monitor 100% of things without some extra configuration, but we give people the ability to do just that with a little work.

tmcdonald wrote:

dlukinski wrote:your general approach to clusters is a simple nodes health plus services running, but not components

The issue with this statement is that, strictly speaking, Nagios XI doesn't really *have* an approach - we can monitor anything as long as there is a plugin for it. If there is a plugin that considers the cluster as a whole then we can absolutely do that. There may well be a plugin on our Exchange that does in fact monitor the cluster as you describe. Otherwise...

dlukinski wrote:we need [...] representation of cluster as a single entity and thus proper capacity/performance reporting.

If a plugin does not currently exist for this, you can use something like BPI (Business Process Intelligence) to logically group hosts/services together into a "meta-group" so to speak, such that if X number of the members are in a non-OK state you can alert based off of that instead of any individual member's status.

https://assets.nagios.com/downloads/nag ... _Addon.pdf

Over the years I have found that Nagios XI monitors about 90% of what people typically need out of the box, another 9% can be monitored using plugins available on our Exchange (or elsewhere), and for about 1% of things there is not a plugin readily-available. I don't know of any software out there that can monitor 100% of things without some extra configuration, but we give people the ability to do just that with a little work.

Maybe we have a misunderstanding as here is what we are looking for:
- say we have active/passive or active/active cluster
- we have to know current and predict future utilization of CPU/RAM for the entire cluster (and not per node, which would not work for active/active scenarios anyways)

Thank you

So you're not talking about just monitoring, it looks like you are specifically referring to the Capacity Planning report? If that is the case then you are correct, currently that only addresses a single host and does not have the logic to aggregate hosts into a cluster as you describe. I would be happy to file a feature request for that if you would like.

tmcdonald wrote:So you're not talking about just monitoring, it looks like you are specifically referring to the Capacity Planning report? If that is the case then you are correct, currently that only addresses a single host and does not have the logic to aggregate hosts into a cluster as you describe. I would be happy to file a feature request for that if you would like.

Please file the feature request.

When it comes to Clusters (windows or linux), how to aggregate CPU/RAM performance stats?

I've created two feature requests with links back to this thread for more information.

Capacity Planning:
XI Aggregation:

You may be able to achieve the aggregation data with this plugin:

https://exchange.nagios.org/directory/P ... 29/details

ssax wrote:I've created two feature requests with links back to this thread for more information.

Capacity Planning:

Code: Select all

NEW TASK ID 7757 created - Nagios XI Feature Request: Capacity Planning - Add the ability to report metrics for an entire cluster (aggregate)

XI Aggregation:

Code: Select all

NEW TASK ID 7759 created - Nagios XI Feature Request: Add the ability to aggregate host/service metrics for clusters

You may be able to achieve the aggregation data with this plugin:

https://exchange.nagios.org/directory/P ... 29/details

Thank you

Please close this case