Nagios-/Icinga-Portal

Hello,
I installed Icinga (Icinga 1.4.0) with Jasperserver (Product Version: 4.0.0 Build: 20110202_0654) and trying to create reports.

On a test system Icinga monitors just localhost.
LOAD for example, if I look at "Service State Trends" in Icinga it says 2.762% up, 0,0% warn, 0,0% unknown, 0,0% crit 97,238% undetermined.
In the Icinga web-interface, Host Availability reports circle diagrams, show that this service is unknown 100%? And other services are OK for 100%.
If I look on the Icinga demo reports there are different states in the circle diagrams, but it doesn't work here.

I also tried to install the 1.5 git version of the reports but that didn't work either.
Does anybody faces the same trouble with the service Availability diagrams?

With kind regards,
Remco Hage

Hi Michael,
Yes there are downtimes in Icinga. At least undetermined in this case (see trends.cgi in attachments).
In Icinga it's 100% unknown or OK (see Schermafdruk.png).

Zitat von »MiCkEy2002«

Hi Remco,

have you Downtimes etc. in your availability report in icinga? The availability report in jasper don't calculate with downtimes. Can you post a screenshot, please?

Cheers
Michael

With kind regards,
Remco Hage

Hello,
Here are 2 screenshots.
One from Icinga Availability Reports and a report from Jasperserver.
Icinga status is 4.488% Unknown and Jasperserver report gives 100% Unknown.
I also tried to read data out of a Nagios (ndo) database (after changing database names) from a larger environment with more data, but all status diagrams are 100% up or unknown. So i looked at the Icinga demo and noticed that it's working correctly there. Therefore I had to rebuild a clean Icinga test system with JasperReports but still facing the same issue.

Hello Michael,
I cannot find errors in the ido2db.

The ido2db.debug log:
[1308299593.277246] [001.2] [pid=646] idomod_broker_data() start
[1308299593.277356] [001.2] [pid=646] idomod_write_to_sink() start
[1308299593.277373] [001.2] [pid=646] idomod_write_to_sink(
211:
1=1200
2=0
3=0
4=1308299593.277242
106=1308298051
102=646
28=1
55=1308299593
60=0
88=1
9=1
97=1
8=1
96=1
39=1
47=1
45=1
103=0
92=0
94=0
78=0
80=0
49=
50=
999

)
[1308299593.277388] [001.2] [pid=646] idomod_sink_buffer_items()
[1308299593.277426] [001.2] [pid=646] idomod_write_to_sink() end
[1308299593.278388] [001.2] [pid=646] idomod_broker_data() end
[1308299593.278456] [001.2] [pid=646] idomod_broker_data() start

I can't see no errors in datasink connection or items flushed.

I looked in phpmyadmin and data is being updated in tabel icinga_servicestatus, see screenshot.
Can I check if data is put in the database correctly?

1. this is the idomod debug log not the ido2db

I attached the ido2db.debug. No errors that i can see.

2. servicestatus isn't read by historical parsers

I was looking there to see if any new data went into the database.
The service availability report querie reads data out of icinga_statehistory as far as i can see in the template. How the querie works i have no idea ( AND state_time between $P{p_date_start} and $P{p_date_start} ?)

select sum(state_duration),
state_one as state,
state_two
FROM
(SELECT 0 AS state_duration,
0 AS state_one,
0 as state_two,
@a:=statehistory_id as counter,
0 as last_time,
0 as last_state
FROM icinga_statehistory
WHERE instance_id = 1
AND state_type = 1
AND object_id = $P{p_service_object_id}
AND state_time between $P{p_date_start} and $P{p_date_start}
UNION
--
-- select all statehistory entries from database and diff them with upcoming entry
--
SELECT TIMESTAMPDIFF(SECOND, IF (sstart.state_time > $P{p_date_start}
AND @A IS NULL
OR sstart.state_time IS NULL, $P{p_date_start}, sstart.state_time), IF (send.state_time > $P{p_date_end}
OR send.state_time IS NULL, $P{p_date_end}, send.state_time)) AS state_duration,
IF(@A IS NULL, IFNULL(
(SELECT state
FROM icinga_statehistory
WHERE instance_id = 1
AND state_type = 1
AND object_id = $P{p_service_object_id}
AND state_time < $P{p_date_start}
ORDER BY state_time DESC LIMIT 1
), 0) , sstart.state) AS state_one,
send.state AS state_two,
@A:=sstart.statehistory_id AS counter,
@B:=send.state_time AS last_time,
@C:=send.state AS last_state
FROM icinga_statehistory AS sstart
INNER JOIN icinga_statehistory AS send
ON send.statehistory_id =
(SELECT MIN(statehistory_id)
FROM icinga_statehistory
WHERE statehistory_id > sstart.statehistory_id
AND instance_id = 1
AND state_type = 1
AND object_id = $P{p_service_object_id}
AND state_time > $P{p_date_start}
)
WHERE sstart.instance_id = 1
AND sstart.state_type = 1
AND sstart.object_id = $P{p_service_object_id}
AND sstart.state_time BETWEEN $P{p_date_start} AND $P{p_date_end}
UNION
--
-- if last existing database entry is smaller then given enddate we have to add the timediff with last state
--
SELECT IF(@B < $P{p_date_end} AND @A IS NOT NULL, TIMESTAMPDIFF(SECOND,@B,$P{p_date_end}), 0) AS state_duration,
@C AS state_one,
0 AS state_two,
@A AS counter,
0 AS last_time,
0 AS last_state
FROM DUAL
UNION
--
-- no entry found in database, simulate an entry for given period
--
SELECT IF(@A IS NULL, TIMESTAMPDIFF(SECOND, $P{p_date_start},$P{p_date_end}), 0) AS state_duration,
IFNULL(
(SELECT state
FROM icinga_statehistory
WHERE instance_id = 1
AND state_type = 1
AND object_id = $P{p_service_object_id}
AND state_time < $P{p_date_start}
ORDER BY state_time DESC LIMIT 1
), 0) AS state_one,
0 AS state_two,
@A AS counter,
0 AS last_time,
0 AS last_state
FROM DUAL
UNION
--
-- delete variable for next run to prevent errors regarding identical database connection
--
SELECT IFNULL(@A:=NULL,0) AS state_duration,
0 AS state_one,
0 AS state_two,
0 AS counter,
0 AS last_time,
0 AS last_state
FROM DUAL
) as state where state_duration > 0
GROUP BY state_one

I posted a screenshot now, off the icinga_statehistory. When a service state changes, it's being updates. I renamed check_load, and after that the change was recorded in the table. Also in my Nagios environment, this table is being filled, so no errors in the log file and new data in the database. Looks like idomod is working to me. Is there a way to check?

The icinga report package I found on http://sourceforge.net/projects/icinga/files/ and followed the manual on https://wiki.icinga.org/display/Dev/Installation+Guide.

Hello Michael,
I've set it to 0 now, but don't see much difference myself.

With kind regards,
Remco Hage

Hello Michael,
I changed the query in icinga.report.sub.service.availability.jrxml.data
Somehow the object_id seems to be a problem.

I tried the following:

Working:
AND object_id = "2216"
AND state_time between "2011-06-01 00:00:00" and "2011-06-01 00:00:00"

Not working:

AND object_id = $P{p_service_object_id}
AND state_time between "2011-06-01 00:00:00" and "2011-06-01 00:00:00"

AND object_id = "$P{p_service_object_id}"
AND state_time between "2011-06-01 00:00:00" and "2011-06-01 00:00:00"

AND object_id = $P{"p_service_object_id"}
AND state_time between "2011-06-01 00:00:00" and "2011-06-01 00:00:00"

I also tried the $P{p_date_start} & $P{p_date_end} this looks ok.

Do you know how i can change the object_id = $P{p_service_object_id} in a working variable?

Zitat von »MiCkEy2002«

Hi,

hmm, i've no idea at the moment where is the mistake

Zitat

I was looking there to see if any new data went into the database.
The service availability report querie reads data out of icinga_statehistory as far as i can see in the template. How the querie works i have no idea ( AND state_time between $P{p_date_start} and $P{p_date_start} ?)

$P{p_date_start} and $P{p_date_start} are parameters from jasper. Jasper fill this parameter with your selected start and end time. You can change the parameters with your start and time and test it on your IDO.

Cheers
Michael

Nevermind michael, somehow it seems to work now. I will check if everything is ok now, but seems ok.

thanks for the help!!!

With kind regards,
Remco Hage