Any ESXi is writing and processing periodic heartbeats to it's VMFS filesystems. When there are issues updating heartbeats due to underlying vSAN storage instability, these heartbeats will timeout and "Lost access to volume" is logged in the vmkernel.log file.
On vSAN-enabled clusters the folders of virtual machines (called "VM Namespace") are using a special form of VMFS filesystems to incorporate required files for virtual machines, like the .vmx configuration file, .vmdk descriptor files, vmware.log files, etc.
These vSAN cluster instability can, but not limited to, caused by any misbehaving disks in the entire vSAN environment (on any node), network latency or instability of the vSAN network, any sort of congestion, etc.