When collecting logs, the collector would hang at 12%, while the CPU on the ZVM is high (around 97% usage).
The Zerto database contains a large amount of events written to it, while the maximum is 200,000. If there are considerably more events, the ZVM is trying to clear those. But it does not manage to clear them out fast enough and gets to a timeout. The operation consumes a large amount of resources with the end-result being unable to collect logs properly or even a ZVM crash.
Upgrading to Zerto 3.5U7 and above will fix the issue.
Alternatively, these tweaks can be added instead (which will take effect after a reboot):
t_eventsMaximumCapacity = 200
t_ZvmEventDilutionIntervalInMs = 21600000