CloudEndure does not write any cache or do any sort of journaling to disk. The agent holds a buffer which is large enough to map all volume's blocks ~250MB in memory.
The agent then acts as a sort of write filter and will replicate changed blocks directly from memory to the replication server. In cases where the data no longer in memory, the agent will read the block from the volume directly. This is the case where you may see backlog in the console. The cause of this is the volume of change is greater than the bandwidth available.