Pandora FMS Server Self-Monitoring Recommendations

These are the most important metrics that must be monitored in the Pandora FMS server.

• Ports, TCP Checks with ports 80, 443, 41121 ( Tentacle ) .
• Services ( tentacle_server, pandora_server, mysql, apache ). It is essential to monitor them through the software agent and with the use of module_condition to alert of their downtime. Example:

module_begin
module_name Service Tentacle
module_type generic_proc
module_exec  service tentacle_serverd status | grep PID | wc -l
module_condition = 0 <script_to_send_email_administrator>
module_end

• CPU. Total CPU usage of the system and the most important services/processes.
• Load AVG.
• RAM memory. Usage of % RAM occupied in the system and usage of each of the required services.
• Disk size occupied in all partitions. With special attention to MySQL datadir .
• Logs of Mysql, Apache and Pandora Server. Alert in case of possible errors.
• Number of .data files in en el directorio /var/spool/pandora/data_in/ and module alert when value exceeds number of agents x2.
• Number of BADXML files in/var/spool/pandora/data_in/.
• Number of Apache, Tentacle and MySQL connections.
• DB Performance (stress). Apply MySQL plugin and monitor its traffic and performance metrics.
• Queued modules and Unknown Agents. These metrics come by default in the server monitoring.

It is important to have defined thresholds and alerts for each of these metrics to keep track of their status and be alerted when a problem occurs in the environment.