In some extreme cases (loss of a datacenter), many proxies send their data at the same time, generating major activity on the base (but on which the indicators remain good) and on the “History syncer” (or DB Syncer, it would be nice to standardize the names of processes and caches).
To absorb these extreme load peaks, it would be nice to be able to increase the cache associated with this process, but this is limited to 2 GB (the cache index settings must obviously follow). 
Wouldn't it be possible to increase the cache's capacity, at least up to 8 GB?