Add retries when there are connections between client VM and allocator VM. Also, add robust logs to indicate the error clearly.