Currently the running total portion of the quota calculation is poorly optimized. Many unnecessary queries are made due to checks happening after the lookup for existing blobs is done. Moving those checks to before the blob lookup will prevent a significant amount of unnecessary queries.