Availibility reports throttle cpu

[WillemDH]

Hello,

Making availibility reports throttle cpu completely. This is just not very practical. While I'm running reports users are not able to do anything in Nagios and avail.cgi uses 100% of cpu. Is there some way to please make this process less cpu hungry?
Even worse, when I want to save the report as pdf, I have the feeling, I need to run the report three times? Once to show the report, then I can choose to create the report as pdf and then when I click the save button, it again takes up to 10 minutes before I can actually save it...
Also I have disabled reports to autorun, but my 'favourite' reports do still seem to autorun...
And it also seem not possible to schedule quarterly reports. (made this feature reqest for that http://tracker.nagios.com/view.php?id=700)

Grtz

Willem

[tmcdonald]

What parameters are you using for the availability? Last year, last month, etc? And is this for all hosts and services or just for a subset?

I will be sure to bring this to the developers, we just need to make sure we know what is triggering this effect.

And how many cores/CPUs do you have on the machine? Is the system info in your signature accurate and up-to-date?

[WillemDH]

Trevor,

I'm querying availibility for some hostgroups containing maximum 20 hosts. To be more precise, I have

- one hostgroup containing our top 10 website
- one hostgroup containing some bpi hosts that contain our Exchange environment
- one hostgroup containing 20 critical network devices
- one hostgroup containg two critical internal websites

For each of the above I've been working all day to get "last quarter" availibility.

I managed to do the above in about 4 hours (which is crazy just to make 4 pdf reports), but as I will have to do this every quarter, I would really like some way to

- make this less intrusive on our production system
- automate this somehow

I'm doing this on a Nagios XI Enterprise edition on R2.6. The system has 6 cpu's at the moment and 8 GB RAM. Load is on average around 1.8. Number of hosts and services in my signature is plus minus correct.

Code: Select all

grep "memory_limit\|max_execution\|max_input" /etc/php.ini
; max_input_time
max_execution_time = 90
max_input_time = 60
;max_input_nesting_level = 64
memory_limit = 512M

Grtz

Willem

[tmcdonald]

Can you run one of the reports and while it is running, provide the output of "top" for us please?

[WillemDH]

Sure Trevor, here you go. Attached one of CPU load and utilisation of last 30 days for you too. I would like to point out that while testing this for you, I go to a saved report (My Reports) and the report started running again without me clicking start.... This while I have the autorun reports disabled...

[tgriep]

How many years of data to you have on this server?

The setting for Disable Autorun of your reports, just prevents the report on the main page from running. If you click on a report, it will run.

[WillemDH]

Tom,

About three years of data. But does that matter when I'm trying to generate reports for the last quarter?

Grtz

Willem

[tgriep]

The 3 years of data shouldn't matter.
When a report is run, it parses through the archived and current nagios log files starting at the start date of the report to the end date.
If you can increase the I/O speed to your drives, that will make it run faster.

[WillemDH]

Tom,

I'm not so worried about a quarterly report taking time. The problem is that the report using 100 % cpu blocks me from using Nagios XI. (and the fact that it takes time makes this worse)

[tgriep]

This is on our feature request list.