Why does wmi.getall fail in the standard Windows OS template by zabbix agent?

Question

Zabbix

Greetings.
Given:
1) 1 subnet like 192.168.0.0/24
2) Windows Server 2019 DC (DNS\AD\SMB)
3) 20 PC Windows 10 Pro (1909) + Zabbix Agent 4.4
4) Debian 10.3 + Zabbix Server 4.4
5) Switch and router (DHCP), but in this context they are not important.

I will say right away that zabbix does not monitor anything critical, it is installed more for collecting statistics and for training.
And now to the main question:
For all hosts with Windows 10 pro (1909), one single template is assigned (Windows OS by zabbix agent). Hosts are queried via DNS, hosts are available, data is coming, graphs are being drawn. BUT! The following lines constantly appear in the logs (go in a row):

613:20200305:204029.598 Zabbix agent item "wmi.getall[root\cimv2,"select * from in32_perfformatteddata_perfdisk_physicaldisk"]" on host "HOSTNAME" failed: first network error, wait for 15 seconds
616:20200305:204044.357 resuming Zabbix agent checks on host "HOSTNAME": connection restored

Sometimes they go for the same host several times in a row. It happens that for 15-20 minutes there is not a single such line in the log, but it happens immediately and for all hosts in a row. Only hosts with Windows 10 are affected, for Windows Server 2019, not a single such line for all time. What is it? Problem with the template for the "civilian" version of Windows? Or is there a configuration issue with me?