Nagios Command File - Timestamp parsing

[Sphinx]

Hey all.

We have installed Nagios Core 3.5.0 and are in the need to process a vast amount from passive check results at once.
The results are comming from a database output and have different check result timestamps.

And here ist the problem, when we insert the passive check result into the nagios.cmd, the leading timestamp is dropped and replaced with the system timestamp when executing.

As an example:
Result Check Result from our test with only one result

Code: Select all

[1425537702] PROCESS_SERVICE_CHECK_RESULT;fkwinbck;Active Backups;0;No Backups Running

Nagios Log at the same second

Code: Select all

[1425541078] EXTERNAL COMMAND: PROCESS_SERVICE_CHECK_RESULT;fkwinbck;Active Backups;0;No Backups Running
[1425541081] PASSIVE SERVICE CHECK: fkwinbck;Active Backups;0;No Backups Running

The CGI's and our database are showing the timestamp of the command execution rather than the check execution.
Is there any chance to configure the command file parser not to overwrite the timestamp?

[Box293]

Can you show us the command you are using to submit the passive check to nagios.cmd

[Sphinx]

Sure, but it's just a redirected output from a file.

Command

Code: Select all

cat passive_check_results >> /usr/local/nagios/var/rw/nagios.cmd

[scottwilkerson]

Not sure what you have in the file but this is the generally accepted way to send results

from http://old.nagios.org/developerinfo/ext ... and_id=115

Code: Select all

#!/bin/sh
# This is a sample shell script showing how you can submit the PROCESS_HOST_CHECK_RESULT command
# to Nagios.  Adjust variables to fit your environment as necessary.

now=`date +%s`
commandfile='/usr/local/nagios/var/rw/nagios.cmd'

/bin/printf "[%lu] PROCESS_HOST_CHECK_RESULT;host1;0;Host appears to be UP\n" $now > $commandfile

[Sphinx]

Hey.

The content of my output is stated on my first post.
The referenced shell script is doing the same thing as mine, basically.
But nagios totally ignores the timestamp in the brackets and replaces it with the current time.

What we need, is to insert check results from days, hours, minutes ago in the command file and want exactly this timestamp and not the timestamp when it is being processed by nagios.

To display it:
the output i want

Code: Select all

[1425537702] PROCESS_SERVICE_CHECK_RESULT;fkwinbck;Active Backups;0;No Backups Running

the output the script produces

Code: Select all

[1425912857] PROCESS_SERVICE_CHECK_RESULT;fkwinbck;Active Backups;0;No Backups Running

What's being inserted into nagios

Code: Select all

[1425912657] EXTERNAL COMMAND: PROCESS_SERVICE_CHECK_RESULT;fkwinbck;Active Backups;0;No Backups Running
[1425912662] PASSIVE SERVICE CHECK: fkwinbck;Active Backups;0;No Backups Running

[scottwilkerson]

What we need, is to insert check results from days, hours, minutes ago in the command file and want exactly this timestamp and not the timestamp when it is being processed by nagios.

This does not work in Core, it cannot write/replace/update the log files into the past.

[Sphinx]

Okay, the timestamp in the log can be the current time, we don't care about that.
But is it possibly to parse our timestamp into nagios event queue instead of the current timestamp, so that we get an entry in the past.
Or are there any other possibilities, without database injection?

[jdalrymple]

... The CGIs get their data from the log, so not really.
There is no database to inject.

If you have NDO then you can use database injection to update the timestamps there, but that won't get used in the Nagios Core interface. Where "in the world" do you need those original check timestamps to be reflected. If in the core web UI... there is no tidy way.

[Sphinx]

Yes, we have an NDO with a MySQL Database.
The use case we have is kind of special and i hoped that it would be possible.

Due to the fact that it is not possible, my question is answered, even if not succesfully resolved.

[abrist]

Fair enough. Nagios Core is tethered to time's arrow . . . .
Lockin' 'er up.