Steps to reproduce:

I'm periodically loose all history and trends data of random VMWare datastore from my hosts. There are 1-2 "loss events" per day. Data losses are random in time and affected host.

It only affects datastore performance data, all other VMWare data are not showing the issue.

Result:

All historical data before the "loss event" is erased in Zabbix (see screenshot).

Whenever such an event happens on any datastore / any host, I see this line in the zabbix_server.log:

9664:20190107:132024.877 Cannot get Datastore info: Timeout was reached.

Expected:

I have enough VMWare collectors enabled and I have increased them from 10 to 30 with a timeout of 30s (was 10 before) each, this did not help. Also no other VMWare performance data but datastore data is affected.

I would expect that - even if the datastore performance could not be fetched once in any minute - that only this certain minute-value would be missing/not available, but under no circumstances this should lead to a complete loss of all previous historical data!