The most convenient way to observe the problem is to execute the "Zabbix stats cluster" item that belongs to the "Zabbix server health" template and check update records for item_tag table in zabbix_server.log (in debug mode). This item triggers the "High availability cluster node discovery" LLD, that in its turn discovers and then checks the following items:
"Cluster node : Address"
"Cluster node : Last access age"
"Cluster node : Las access time"
As result, each time a value is returned for the master item ("Zabbix stats cluster"), item tags of the corresponding discovered items are updated:
Currently this is not a big problem, but once
ZBXNEXT-7554 gets merged, this will add multiple useless records in the changelog table.
Steps to reproduce:
- Open the default zabbix installation
- Open the list of items of Zabbix server host
- Execute item "Zabbix stats cluster"
- Make sure that "Cluster node ..." items are discovered
- Execute the master item again
- Check Zabbix server log (DebugLevel = 4)
Result: each time the master item is triggered, discovered item tags are updated
Expected: discovered item tags should be updated only when something is actually changed.
Server log: zabbix_server.log