Page MenuHomePhabricator

Determine cost-benefit of doing vertical data slicing on WDQS
Closed, ResolvedPublic

Description

As a product manager, I want to understand each vertical data slice and the associated cost and benefits, so that I can be better informed to make decisions in the case of catastrophic failure.

I want to know for each possible vertical slice of Wikidata items (i.e. description, labels, etc), the potential benefit of removing by number of triples/disk size vs the cost in terms of user impact (i.e. number of queries, users affected).

Revisit the vertical analysis of Wikidata (https://wikitech.wikimedia.org/wiki/User:AKhatun/Wikidata_Vertical_Analysis), and provide number of affected queries for each potential vertical slice.