Clickhouse Nodes out of sync causing Code 242 DB:Exception and Keeper_exception

Hello,

I have migrated my FortiSiem DB into clickhouse the last 6 months and I face the following issues.
Firstly, let me clear that I contact the vendor support but they can't resolve the issue apart from telling me to remove keep only 1 keeper node instead of 3.

My architecture: 1 supervisor, 3 keeper only nodes, 2 data nodes x 2 shards (4 total data & query nodes)

All nodes are VMs running on ESXi host. The clickhouse runs properly apart from some times, which is usually once per week, throws this exception in the logs:

**Code 242** DB: Exception, table is in readonly mode
**OR** **Code: 999**, Session expired

Keep in mind that VMs run on the same host and only a VSwitch exist between them for sending/receiving packets.
After investigation in the keeper logs, we discovered possible connectivity issue between keepers causing replicas on data nodes to go in read only state:

keeper2-data-clickhouse-hot-1/clickhouse-keeper/app_logs/clickhouse-keeper.log:2024.05.21 17:23:41.008101 [ 2268 ] {} RaftInstance: Election timeout, initiate leader election
keeper3-data-clickhouse-hot-1/clickhouse-keeper/app_logs/clickhouse-keeper.log:2024.05.20 17:04:09.139589 [ 2253 ] {} RaftInstance: Election timeout, initiate leader election

There are cases where timeout occurs because response time is <500ms

Although I suspect the issue might be a connectivity, I checked the ESXi host that has the VMs running and I cannot conclude that this is the cause because it has low cpu/ram/disk usage.

My questions to the community are:
1) Based on your experience, what are the possible causes of these issues?
2) How can I check that ESXi has issues on CPU wait time for the VMs and if this is the cause, how to improve that?