Hi Bernardo,
Are you usign…
Encode runlength and encode raw and
DISTSTYLE KEY
DISTKEY (root_id)
SORTKEY (root_tstamp) on create tables ?
There is a doc with 10 best pratices on redshift queryng.
But there is no doc for best pratices with “maintenance” on redshift. Search about “WLM dinamic memory allocation” there you can change and create rules for query allocation memory…
– query_concurrency
– memory_percent_to_use
– scan_row_count
http://docs.aws.amazon.com/redshift/latest/dg/cm-c-wlm-dynamic-memory-allocation.html
We run a vaccum on all tables on Redshift… i know that it’is not only the best pratices for betther performance…but… we run in all conext tables and after on atomic.events and others low perfomance tables