Type: Incident report
Affects Version/s: 3.4.13
Fix Version/s: None
We experienced an incident last week end with following symptoms :
History Cache fall down from 100 % to 0, back to ok status after a server restart.
We know this could be linked to an items overflood, so after search, we found an items who produce more than 580 NVPS during the problem.
This item (calculated item) as sent a lot of same value for the same timestamp (clock), but all ns values are differents in the history table.
It appear that after the zabbix_server restart, the problem just stopped.
It's actually not reproducable, but this is not the first time that we had this kind of problem, but it was located on proxy side with ZBX_TCP_READ error in the log file.
We have lost the server log file from an unknow cause during this incident, so, very hard to search for errors in the log file.
Please find in attachment some screenshot regarding the duplicated values.