Drop the spent DeadAir clean-up and keep not-due skips off live updates (#196, #197)

The clean-up of DeadAir's old copies probed for its table on every open, and on
Postgres each failed probe is an ERROR in the server's log: 120 an hour from the
healthcheck alone. Production ran it already and is the only database.

A scan sent a not-due skip for each of ~1,535 feeds into a broadcast of 1,024,
so a page slow to read lagged and dropped events: 443 counted in five hours.
Nothing listening uses them; the terminal and the log still get them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -23,6 +23,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- The database server no longer logs an error each time the healthcheck runs, about 2,900 a day, from a clear-out of DeadAir's old copies that had already run (#196).
+- A page left open no longer misses updates after a check for new episodes: the note that each feed not due was skipped, one per feed, filled the queue they share (#197).
 - Checking for new episodes no longer waits up to 12 seconds each time on a clear-out of dead feeds from the Directory, which now runs once a day (#193).
 - All Subscriptions, and every item list, comes up faster, most of all the first time after a quiet spell: it took half a second to two (#188).