…idStream against CI flakes

Three changes to eliminate the timing-sensitive flake:

1. Change PendingChatAcquireInterval from time.Hour to time.Second.
   The primary trigger is still signalWake() from SendMessage, but a
   short fallback poll ensures the worker picks up the pending chat
   even under heavy CI goroutine scheduling contention.

2. Increase context timeout from WaitLong (25s) to WaitSuperLong (60s).
   The worker pipeline (model resolution, message loading, LLM call)
   involves multiple DB round-trips that can be slow when the
   PostgreSQL instance is shared with many parallel test packages.

3. Add a status-polling loop while waiting for the streaming request.
   If the worker errors out during chat processing, the test now
   fails immediately with the error status and message instead of
   silently timing out after 60 seconds.

Fixes coder/internal#1455