fix: bad pattern matches on reducers in ingest event queue

[amokan]

I found this gap while going through PR #3088 and PR #3089 again this morning.

So in PR #3069 - we introduced IngestEventQueue.list_counts_with_tids/1 and some reducer patterns that used the results from that function like {{sid_bid_pid, tid}, size}.
The next day in PR #3065 - we changed the return of list_counts_with_tids/1 to {table_key, size}, but did not update the reducer patterns. I then built upon those reducer patterns for the consolidated pipeline efforts.

This PR aligns the reducer patterns with the current return shape of {table_key, size}

[Ziinc]

i don't think there is a test that catches this code path, would be good to add that in @amokan , especially since #3088 seems to be passing