![]() The possibility to monitor time deviation using the standard active-mode agent is another very important reason. I understand that it could cause to some problems (for example, false positives if some values have been buffered and delayed by active-mode agents and intermediate proxies), but these problems much more predictable and understandable than the current behaviour. In result, we could use this metric by active-mode agents, too. I believe that the behaviour of this trigger function should be modified according to documentation. Is SHOULD use, or at least allow as a parameter, the TIME THE ITEM WAS COLLECTED, which is known and stored, but not used by fuzzytime. It compares the time reported by the host to the time on the Zabbix server AT THE TIME TRIGGERS ARE PROCESSED. Key moment: "differs from the Zabbix server time" but anything else (including value's timestamp)! And native HTTP agent will discover all proxies and create lastaccess item per each proxy + fuzzytime trigger. We use the 'fuzzytime' function generate an alert if the time of a monitored server has drifted. As result, it makes useless trigger expressions like "host:(SomeValue)" if this metric is collected by agent in active mode, as value and timestamp should be equal "by design" in this case.Īt the same time, documentation always has the following description for trigger function fuzzytime():Ĭhecking how much an item value (as timestamp) differs from the Zabbix server time. The key moment is: the current value of item is compared with timestamp of this value. ![]() ![]() I found that out while going to latest data, checking the timestamp and converting it to actual time.Cyclone in this comment describes how Zabbix server processes the fuzzytime() function at the moment. When I configured the script to return a timestamp, I changed it to a different timezone instead of living it as it is, which actually compared the data with current time + 2 hours in the future :) so besides actually getting the triggers to work properly everything else is well configured. (In theory an isolated server should not be able to send notifications. the problem is that it returns a critical error. Zabbix Server connectivity issues Template App Zabbix Server Zabbix Server is network isolated Template App Zabbix Proxy No connection from proxy Make proxy trigger dependent on this trigger in order to avoid misleading notifications. PGPASSWORD= -U -t -c "" | awk '=0 so if the last insert time is greater then 5 minutes to return a critical error. I perform the query and send the output to zabbix_sender with the following template: #!/bin/bash Ok so first thing I did is created a bash files at /etc/zabbix/mytools, get-last-insert-time.sh. For the key value enter, eventlog Security,4625,skip Note : The skip option for the mode flag at the end. When you sort agent part, then create template with item key and Numeric (unsigned) type. This allows the Zabbix agent to read the windows event logs. it's probably not cause it's not working :) Test it with zabbix-get from the Zabbix server, e.g.: zabbixget -s -p 10050 -k '' For any issue on the agent side: increase log agent level and watch agent logs.all I did so far is for googling, not sure if that's the way to go. If the last inserted row have a insert time of more then 5 minutes to product warning and 10 minutes to product error. ![]() I have a a table that I want to check a timestamp column for the last inserted row. I have a postgresql plugin configured and working properly so it checks my database metrics. I have Zabbix 4.4.1 installed on Ubuntu 19.10.
0 Comments
Leave a Reply. |