No retry logic for transient failures in services

[user1303836]

Problem

Most services lack retry logic for transient failures. When external services have temporary issues (network hiccups, rate limits, 5xx errors), operations fail permanently instead of retrying.

Affected Services:

Service	Issue
content_extractor.py	No retry on network errors
page_analyzer.py	Retries RateLimitError only, not timeouts or 5xx
content_poster.py	No retry on Discord rate limits or temporary unavailability
message_forwarder.py	No retry on Discord API hiccups
summarizer.py	Retries RateLimitError but not other transient API errors

Impact

• Temporary Discord outages cause messages to never be posted
• Network hiccups cause content to be skipped permanently
• Items that fail once are never retried (content_poster marks as "attempted" but item remains unposted)
• Users lose content due to transient issues

Proposed Solution

1. Add retry decorators with exponential backoff for transient errors:

from tenacity import retry, stop_after_attempt, wait_exponential, retry_if_exception_type

@retry(
    retry=retry_if_exception_type((httpx.TimeoutException, httpx.HTTPStatusError)),
    wait=wait_exponential(multiplier=1, min=2, max=30),
    stop=stop_after_attempt(3),
)
async def fetch_content(self, url: str) -> str:
    # ...

2. Distinguish between transient errors (retry) and permanent errors (fail immediately):

   • Transient: 429, 500-599, timeouts, connection errors
   • Permanent: 400, 401, 403, 404

3. Implement a "dead letter queue" for items that fail after max retries

Labels

enhancement, reliability