One frequent and strange bug we are facing with firestore is the dropping of events when designing large batch onWrite-listeners. This leads to unexpected bugs in production code.

For some reason, when executing say 300 to 1000 onCreates at once, different events seem to drop and are never responded too.

A way to access the tracelog of firestore events (in Stage 1 Cloud functions) would help us debug this problem. It would also help replay event streams in case the database is corrupted due to non-responding to the event streams.