support request with sql in apiserver

It's better for debug test. We can't test deployment quickly(needs table, index, data, and create/drop deployment), so run request with sql is lightweight

Look at the apiserver code to understand how SQL requests are currently handled. The feature likely involves adding a new endpoint or modifying an existing one to accept raw SQL for debugging. Check for existing test deployment procedures to see what tables, indexes, and data are needed. Determine how to make this lightweight and safe for debugging purposes.

索引モデルが issue の本文から書いたものです。