I have a ruby script that runs some simple statistics via a MySQL connection, and on some larger databases, they may take ~10 seconds, but assume that under heavy load, they can take a minute or so. Here's my predicament as it currently stands (and please call me out if anything here is incorrect):
If I use a UserParameter and zabbix_get, I get "ZBX_NOTSUPPORTED" on the long ones. The #zabbix channel on Freenode suggested that I use zabbix_sender to push the results directly to a trapper, however...
If I use a sender/trapper, I can't seem to create universally-unique templates. On the server (via the web interface), I need to have a unique identifier in "Key", and it apparently doesn't take macros, so this means a couple dozen items across about a hundred agents. Not maintainable
Any suggestion on what to do here?
If I use a UserParameter and zabbix_get, I get "ZBX_NOTSUPPORTED" on the long ones. The #zabbix channel on Freenode suggested that I use zabbix_sender to push the results directly to a trapper, however...
If I use a sender/trapper, I can't seem to create universally-unique templates. On the server (via the web interface), I need to have a unique identifier in "Key", and it apparently doesn't take macros, so this means a couple dozen items across about a hundred agents. Not maintainable

Any suggestion on what to do here?

Comment