1. Register vmware.eventlog monitoring(vmware.eventlog[{$URL},skip])
2. ESXi server shutdown.
3. Items in vmware.eventlog become unreachable (Couldn't connect to server)
4. ESXi server restored.
5. After one day, vmware.eventlog is not recovered. (Other monitoring, such as vmware.hv.memory.used, will be restored.)

If you check the debug log, you will see that the vmware collector acquires 10 eventlogs at a time at the acquisition timing.

1906315:20210527:113047.135 In vmware_service_parse_event_data() last_key:589181
1906315:20210527:113047.135 skipping event key '172214', has been processed
1906315:20210527:113047.135 skipping event key '172213', has been processed
1906315:20210527:113047.135 skipping event key '172212', has been processed
1906315:20210527:113047.135 skipping event key '172211', has been processed
1906315:20210527:113047.135 skipping event key '172210', has been processed
1906315:20210527:113047.135 skipping event key '172209', has been processed
1906315:20210527:113047.135 skipping event key '172208', has been processed
1906315:20210527:113047.135 skipping event key '172207', has been processed
1906315:20210527:113047.135 skipping event key '172206', has been processed
1906315:20210527:113047.135 skipping event key '172205', has been processed

This situation continues until the lastlogsize of the item_rtdata table, so depending on the situation, the "Couldn't connect to server" message may continue for a week or more.

Zabbix server goes down and recovers, I think the behavior is to get up to 1000 event logs.
If ESXi recovers and the message "skipping event key 'xxxx', has been processed" output in the log, is it possible to retrieve up to 1000 items?

Also, since the connection to ESXi is established, I think it would be easier to understand the status by updating the "Couldn't connect to server" message.