As someone working on WDQS, I want to better understand the limits of the service, the use cases that we don't support but that are consuming resources.
Looking at queries that timeout but could provide results should give us some idea of what users are trying to achieve, but are not able with the current limitations of the service. Since those queries are by definition expensive, it might also give us some ideas on how to reduce the load on the servers.
Things to do:
- extract a random subset of queries that are in timeout
- analyze this subset to understand what kind of queries this is and if we should be doing something about it