The archive_results_db sink is the long-term home of every search
result. It writes to the shared Postgres server and keeps a larger connection
pool than the hot database, because every result lands there while only
high-match results reach the hot one.
When this sink reports delivery errors, check the Postgres service first: the pipeline keeps consuming, and failed batches are not replayed, so a long outage is visible as a gap in the archive rather than as a stalled worker.
search-requests is flat or falling.