Hi all.
I am running a zabbix v6.0 server and I am think about monitoring some logfiles on debian server with zabbix-agent2 installed.
I need to create N log items that work on the same logfile but with different regexp.
Is this a good practice?
I thought about a different way to implement this, with a common "base" item
and then get the different keys with N dependant items with a "regexp" preprocessing rule.
I suppose the first approach puts more load on the agent while the second one puts more load on the server.
How much? Do considerations about this vary depending on N (<10, >10, >50)?
What about if I need only .count ?
What would you advise?
I am running a zabbix v6.0 server and I am think about monitoring some logfiles on debian server with zabbix-agent2 installed.
I need to create N log items that work on the same logfile but with different regexp.
Code:
custom.key1=logrt[/var/log/logfile.log,regexp1] custom.key2=logrt[/var/log/logfile.log,regexp2] ... custom.keyN=logrt[/var/log/logfile.log,regexpN]
I thought about a different way to implement this, with a common "base" item
Code:
custom.basekey=logrt[/var/log/logfile.log,regexp1|regexp2|...|regexpN]
I suppose the first approach puts more load on the agent while the second one puts more load on the server.
How much? Do considerations about this vary depending on N (<10, >10, >50)?
What about if I need only .count ?
What would you advise?