Nagios Support Forum

One of the reasons we are upgrading to xi from Core, is to maintain and report on historical client performance data. I'm doing the xi install now, but need to quickly monitor a few clients for load, memory, etc and produce a report. Some questions come to mind:

Once monitoring a service on the client (performance metric, like load, disk space, #processes, etc), do I have to do anything special to store the values rather than have them discarded like in Core?

I believe the stored values are kept in MySQL, right? Can I get a Data Dictionary (if that's the right term) for the database, so that we can generate our own reports with Crystal Reports? Our reports person says we already have an ODBC Connector for MySQL, so just need the Data Dictionary to start doing queries and reports. It would be great if someday xi could generate reports on performance metrics like this.

I'm not too familiar with RRD Tool, but I believe that's a component of xi. Do I have to take any extra steps to install this? Will RRD Tool automatically "digest" the client performance data within MySQL, so that Crystal Reports can report on load (for instance) from yesterday, last week, last month, etc?

Can you point me to something in the xi docs I should be reading for this area?

Thanks for any advice....Lyle

Once monitoring a service on the client (performance metric, like load, disk space, #processes, etc), do I have to do anything special to store the values rather than have them discarded like in Core?

No, performance data is retained and graphed by default.

I believe the stored values are kept in MySQL, right? Can I get a Data Dictionary (if that's the right term) for the database, so that we can generate our own reports with Crystal Reports?

I'm not familiar with that terminology, but I've attached the SQL description.

It would be great if someday xi could generate reports on performance metrics like this.

We are significantly retooling the entire reporting framework for a new upcoming major release. Feel free to provide details on exactly what sorts of reports are useful to you and we'll take them into consideration with other suggestions for designing that.

I'm not too familiar with RRD Tool, but I believe that's a component of xi. Do I have to take any extra steps to install this? Will RRD Tool automatically "digest" the client performance data within MySQL, so that Crystal Reports can report on load (for instance) from yesterday, last week, last month, etc?

Yes, RRDTool is part of xi, and installed as part of the standard installation process through the provided scripts. It is used to generate the graphs of performance data shown in the xi interface. As for your Crystal Reports integration, you would probably be better off fetching the database entry selecting by a timestamp range.

Can you point me to something in the xi docs I should be reading for this area?

Assuming you understand SQL, the attached file will certainly tell you a lot about what is available. Beyond that I'm not sure we have anything available of the sort you might be looking for, but if you could provide any further details of what you need after reading this post I'll see what I can do.

Thanks, Tony. I've passed the "SQL description" on to out Crystal Reports person to see if that's what they are looking for.

tonyyarusso wrote: No, performance data is retained and graphed by default.

If I've run a check_nrpe command to get the load on a client, and that load value is checked against the warning & critical levels, it's value is then thrown away? That's not what I understood in a pre-sales call. I understood there was an add-on packaged with xi that retained the data.

Perhaps I'm thinking of PNP. From the PNP website: "PNP is an addon to nagios which analyzes performance data provided by plugins and stores them automatically into RRD-databases"
This also suggests the RRD involvement.

Thanks once again. But now I'm nervous. We *need* to retain historical performance data so we can report on trends.

...Lyle

From the "Nagios xi Architecture" page:
"PNP integration provides automatic, RRD-based performance and trending graphs"

Thanks...Lyle

Punctuation is critical there - note the comma. To put it more clearly, "No, you do not have to do anything special to store the values, as performance data is retained and graphed by default."

I've now got xi installed and monitoring a couple of our servers, so now I want to ensure that metrics collected by the services are being stored, and can be accessed by our reports person, come Monday.

Keeping it simple, one service checks the total number of processes running on the client. nrpe runs the check_procs plugin on the client, and the xi screen reports:
"PROCS OK: 94 processes"

Where is the number 94 stored? in MySQL?

How can I use xi to see or graph this value and it's predecessors?

I have the MySQL database definition that you've provided, but would this definition describe how to get to this value and it's predecessors with an SQL query?

I don't really understand how PNP, RRDtool, and Cacti are involved with xi. Is there any documentation, other than that on the individual product websites?

Thanks...Lyle

The metrics for each check (e.g. "94 processes") are not stored historically in the Nagios xi databases. If the individual plugin returns performance metrics in a manner that can be graphed by PNP (e.g you see a performance graph in Nagios xi), the the data for the processes will be stored in an underlying RRDtool datafile.

The RRD files for all hosts/services that have graphs can be found in the following directory on the xi server:
You'll see subdirectories that contain performance data in an RRD file (.rrd extension).

You could use the rrdtool command to get data out of the file. The usage instructions for doing this can be found in the RRDTool documentation.

Ah, sorry - the particular table I was thinking would have it only records the output when the state changes. What admin said about RRD is what you will need. In xi there is a tab on each host and service labeled "Performance Graphs" where graphs of this data can be viewed, as well as options for viewing all graphs or all graphs for a particular host.