====== PoolRotater v6.4 — Outage Recovery & Multiroom Stability ======

**Historical archive — not current configuration guidance.** Source: Papapyrus's Discord announcement, supplied by the project owner for archival publication. Posted: September 10, 2026, 9:40 PM. Displayed time zone was not supplied. Announcement timestamps are not independently verified deployment dates.

===== Announced changes =====

A reliability update for network, Discord and osu! service interruptions.

  * Detects broken/stale shared referee connections and rebuilds after disconnects, DNS failures or socket loss.
  * Active rooms rejoin the referee connection after recovery rather than using stale membership.
  * Unknown referee outcomes are handled conservatively instead of blindly retrying match starts, playlist changes or room closure.
  * Pending referee commands wake immediately when the connection dies instead of waiting on long timeouts.
  * Heartbeat/watchdog checks detect connections that appear alive but have stopped receiving data.
  * During chat reconciliation, live messages wait until missing history has been checked instead of advancing across a possible gap.
  * Reworked Ctrl+C shutdown handling avoids indefinite hangs while closing unreachable rooms.
  * Failed cleanup reports affected room IDs and the correct manual cleanup command rather than retrying forever.
  * Discord cards and channel-name indicators run independently, so rate-limited renames do not block cards or replies.
  * Coalesces Discord status updates to reduce API traffic while keeping open/closed transitions responsive.
  * Lobby names hot-load from JSON: immediate changes when safe, deferred until after starting/playing maps.
  * Adds configurable ''playlist_audit_seconds'' to slow room-state safety checks without removing pre-match reconciliation.
  * ''playlist_audit_seconds'' can change live through ''universal.json'' without recreating rooms.
  * Expands regression coverage for DNS failures, dead sockets, room rejoin, stuck shutdowns and multiroom isolation.

Goal: make failures recoverable rather than catastrophic. A temporary outage should not take down the whole multiroom setup because the shared referee connection became unhealthy. Further API-usage reduction while retaining safeguards was announced as the next focus.

For current behavior, use [[wiki:start|the feature guides]] and [[commands:start|the command reference]]. [[updates:start|Back to release history]].