Resume retries after cooldown instead of giving up for the day

Previously 3 failed attempts at 06:00 (e.g. llamaswap model cold/starting)
permanently blocked generation until the next day. Now the retry loop
pauses SLOW_RETRY_MINUTES (default 60) after the burst budget is spent,
then resumes — one attempt per hour until jokes exist.

.env.example

@@ -9,9 +9,12 @@ OPENAI_API_KEY=***
 GENERATE_HOUR=6
 GENERATE_MINUTE=0
 
-# Failure handling: retry every RETRY_MINUTES, up to MAX_ATTEMPTS per day
+# Failure handling: retry every RETRY_MINUTES, up to MAX_ATTEMPTS per day.
+# After that, retries pause for SLOW_RETRY_MINUTES, then resume (so a cold
+# LLM endpoint at 06:00 doesn't kill the whole day).
 MAX_ATTEMPTS=3
 RETRY_MINUTES=15
+SLOW_RETRY_MINUTES=60
 
 # Wikipedia language for the featured-article feed
 WIKIPEDIA_LANG=en