hi Alexi,
Just some little things comming to my mind:
1/ it could be handy to have a "test now" button for items so you can ask in real time a probe. Can be usefull for items with refresh rate that are 86400 for exemple
2/ capability to change configuration file path for agent. i tried for server it's ok but for agent i cannot (i also want to put them in /usr/local/etc/zabbix on freeBSD).
3/ double check. React to a trigger by issuing a new poll in a defined interval. Launch action only in case both fail. Exemple i check a icmpping and it fail, i have configured doublecheck(3), three second later it issue another icmpping. If it fails also then Action is triggered, if not then it's considered normal network glintch or perhaps trigger another action (issue just a warning instead of an alerts).
This is mainly to compensate for single server alert system where network glitch can throw false alerts. This is different from testing twice the same item and react only to a 2 check failure in the fact that you double the probe only in case of failure, this bring much more scalability for the server. I do not think actions can trigger a "trigger" right now ?
4/ simple check additions, http_value(regex) return the value found by the regex if any. There you can trigger on the return value easely. This will prevent the use of a local agentd that itself connect using wget and pipes everywhere. this also allow to monitor some very simple things without having to install a zabbix agent on the host (simple php script).
5/ DNS simplecheck, checkdns(MX,ns1.hotmail.com), return string of records séparated by ";" or FAIL if no answer. This is usefull when you monitor an internet host to also check dns services
best regards,
Ghislain.
Just some little things comming to my mind:
1/ it could be handy to have a "test now" button for items so you can ask in real time a probe. Can be usefull for items with refresh rate that are 86400 for exemple
2/ capability to change configuration file path for agent. i tried for server it's ok but for agent i cannot (i also want to put them in /usr/local/etc/zabbix on freeBSD).
3/ double check. React to a trigger by issuing a new poll in a defined interval. Launch action only in case both fail. Exemple i check a icmpping and it fail, i have configured doublecheck(3), three second later it issue another icmpping. If it fails also then Action is triggered, if not then it's considered normal network glintch or perhaps trigger another action (issue just a warning instead of an alerts).
This is mainly to compensate for single server alert system where network glitch can throw false alerts. This is different from testing twice the same item and react only to a 2 check failure in the fact that you double the probe only in case of failure, this bring much more scalability for the server. I do not think actions can trigger a "trigger" right now ?
4/ simple check additions, http_value(regex) return the value found by the regex if any. There you can trigger on the return value easely. This will prevent the use of a local agentd that itself connect using wget and pipes everywhere. this also allow to monitor some very simple things without having to install a zabbix agent on the host (simple php script).
5/ DNS simplecheck, checkdns(MX,ns1.hotmail.com), return string of records séparated by ";" or FAIL if no answer. This is usefull when you monitor an internet host to also check dns services
best regards,
Ghislain.
Comment