Audience asks whether events should simply go in a database, or in memory/logs, etc. Ismael replies that despite years exploring event sourcing he has virtually zero production experience, so he can't give authoritative advice; he started on Postgres and switched to SQLite to simplify dependencies and focus on patterns. He notes others use event sourcing at scale on Postgres or even files. An audience member from the floor confirms Postgres works really well for it in production.