Using api/issues/search to load all issues is a too big risk. The WS is limited to 10k issues, and returns much more data than we need, which may put too much load on the server.
Instead we will do an hybrid strategy:
- use batch/issues to load all issues, as before
- if we find some taint vulnerabilities in the response, we will call api/issues/search to load missing data (precise location and flows)