Before the update, 6sec freeze had been sufficient on my system.
In the last run (9.5U3) the timescale is as follows:
13:41:05 VSS freeze started
13:41:07 Create virtual machine snapshot - Started
13:41:08 Create virtual machine snapshot - Completed
13:41:25 Shadow copy freeze ended (aborted)
As far as I understand the following step causes the problem (KB1680):
Return of snapshot information via VIM API
These steps should usually be near-instantaneous, but if the vCenter is heavily loaded or has a high latency to the ESXi hosts, the delay may be significant.
The vCenter is the same, I blame Veeam for this issue