Monitoring the reporting server¶
In order to garantee the availability of Centreon Monitoring Business Intelligence and it’s services, you will need to monitor the reporting server by adding manually some indocators, or use the plugins packs system to add easily the indispensable indicators.
Method 1: Monitoring using Centreon plugin packs¶
The plugin pack centreon-pack-Applications-Monitoring-Centreon-Mbi allows to monitor the critical process and the state of the reporting server. To install it, launch the following command on your monitoring platform ( Central server )
yum install centreon-pack-Applications-Monitoring-Centreon-Mbi
The plugin pack is now installed, just follow the procedure included in the plugin pack. We advice you to also link your reporting server to App-DB-MySQL template. (You’ll need to create a MySQL user to use that plugin pack)
Method 2: Adding the monitoring manually¶
If you didn’t purchase yet Centreon plugins packs , you will find below all indicators to add in order to monitor correctly your reporting server:
The following process must be running on the reporting server.
|mysqld||Database server hosting Centreon MBI datawarehouse|
|cbis||Job scheduler for the report generation|
Here is an example of a check command to monitor mysqld process and its memory consumption:
$USER1$/centreon_plugins.pl --plugin=os::linux::snmp::plugin --mode=processcount --hostname=$IP_SERVEUR_REPORTING$ --snmp-version=$SNMP_VERSION$ --snmp-community=$SNMP_COMMUNITY$ --process-status --process-name='^mysqld$' --regexp-name --regexp-path --regexp-args --critical='1:1' --memory
Datawarehouse integrity monitoring¶
A plugin is provided with Centreon MBI that you can use to check :
- if the content of the database is up to date
- the number of jobs in “failed” state
Data integrity check (execute it by default 1 times a day) :
The plugin returns CRITICAL and displays the name and the last entry of the tables which are not up to date. This check take fiew seconds. Set the timeout of the check to 30 seconds or adjust it if necessary.
Number of failed jobs :
The plugin return the number of jobs in failed states and a return a CRITICAL states if > 0
Server capacity monitoring¶
The following checks will help to anticipate the evolution of the reporting server capacity:
- The load average
- The CPU consumption (by core and the average)
- The RAM (independently of the SWAP)
- The SWAP (independently of the RAM)
- The storage space
- The I/O disk
ETL execution logs monitoring¶
The following file hosted on the reporting server should be monitored:
If a pattern FATAL or ERROR appears in one of these files, it means that the execution went wrong.