A failing feed is retried on its normal schedule forever #99
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
A feed that fails is checked again at its usual interval (every 1h by default), however long it has been failing: gizmodo (404), pelgrane-press-ltd (403), daily-quests (503) and toddstashwick (redirect loop) each fail every hour, all day (feed_error events, 2026-09-29). It costs a request to a site that has said no, a WARN line each time, and scan time. Wanted: exponential backoff, the wait growing the longer a feed has been failing, up to a ceiling, reset by the first success; a refresh someone asks for still tries at once.
Fixed in
db87bc0: a failing feed waits as long as it has been failing (error_since to its last check), never less than its usual interval, never more than a day; 1h, 1h, 2h, 4h, 8h, 16h, then daily on an hourly schedule. The first success resets it; a refresh still tries at once. After the deploy, daily-quests, gizmodo and toddstashwick (failing since 2026-09-14) are next tried 24h after their last try. Deployed.