Split the work into threads
At least in the debug mode the query can take some time. We don't want to stop accepting other queries or the reports from guts
. This would get even worse once we store the data on-disk.
We should introduce threads into the processing. There are some specialized threads we might want to start, eg: for updating the keeper data.
Furthermore, we want to submit the queries into worker threads, eg futures-cpupool
.