Nodata trigger function not working as intended
Hello,
After migrating from Zabbix 5.0.x to 6.0.y we are having issues with "nodata" triggers.
Documentation https://www.zabbix.com/documentation...ctions/history states that:
Since Zabbix 5.0, the 'nodata' triggers monitored by proxy are, by default, sensitive to proxy availability - if proxy becomes unavailable, the 'nodata' triggers will not fire immediately after a restored connection, but will skip the data for the delayed period. ... For active proxies suppression is activated if connection is restored more than 15 seconds later.
When we used Zabbix 5 (5.0.30 for example), suppression used to work correctly: if a Zabbix-proxy is down/unavailable etc., triggers with nodata were silenced and only triggers with "strict" option would fire. After migrating to Zabbix 6 (currently 6.0.12), just after proxy becomes available, ALL triggers with nodata function fire immediately for a few seconds. This causes hundreds of events and notifications for us. Were any changes done that were not documented regarding this function?
Zabbix server is running on AlmaLinux 9.1, version 6.0.12.
Proxies are in active mode, on CentOS 7.9, version 6.0.12.
Trigger example:
nodata(/serverX/vfs.fs.size[C:,free],600s)=1 and last(/serverX/system.uptime)>10m and nodata(/serverX/agent.ping,15m)=0
Recent event example:
Problem created: 2023-02-10 12:06:37
Problem resolved: 2023-02-10 12:07:35
From proxy log, in this case it was a short network problem between proxy and server from 2023-02-10 12:05:50 to 2023-02-10 12:06:10:
30100:20230210:120550.049 Unable to connect to [xxx]:10051 [cannot connect to [[xxx]:10051]: [4] Interrupted system call]
30100:20230210:120550.049 Will try to reconnect every 1 second(s)
30100:20230210:120610.059 Connection restored.
Any ideas why suppression not working since Zabbix 6 or am I missing something in the documentation?
After migrating from Zabbix 5.0.x to 6.0.y we are having issues with "nodata" triggers.
Documentation https://www.zabbix.com/documentation...ctions/history states that:
Since Zabbix 5.0, the 'nodata' triggers monitored by proxy are, by default, sensitive to proxy availability - if proxy becomes unavailable, the 'nodata' triggers will not fire immediately after a restored connection, but will skip the data for the delayed period. ... For active proxies suppression is activated if connection is restored more than 15 seconds later.
When we used Zabbix 5 (5.0.30 for example), suppression used to work correctly: if a Zabbix-proxy is down/unavailable etc., triggers with nodata were silenced and only triggers with "strict" option would fire. After migrating to Zabbix 6 (currently 6.0.12), just after proxy becomes available, ALL triggers with nodata function fire immediately for a few seconds. This causes hundreds of events and notifications for us. Were any changes done that were not documented regarding this function?
Zabbix server is running on AlmaLinux 9.1, version 6.0.12.
Proxies are in active mode, on CentOS 7.9, version 6.0.12.
Trigger example:
nodata(/serverX/vfs.fs.size[C:,free],600s)=1 and last(/serverX/system.uptime)>10m and nodata(/serverX/agent.ping,15m)=0
Recent event example:
Problem created: 2023-02-10 12:06:37
Problem resolved: 2023-02-10 12:07:35
From proxy log, in this case it was a short network problem between proxy and server from 2023-02-10 12:05:50 to 2023-02-10 12:06:10:
30100:20230210:120550.049 Unable to connect to [xxx]:10051 [cannot connect to [[xxx]:10051]: [4] Interrupted system call]
30100:20230210:120550.049 Will try to reconnect every 1 second(s)
30100:20230210:120610.059 Connection restored.
Any ideas why suppression not working since Zabbix 6 or am I missing something in the documentation?
Comment