Problem
Library Manager paces BookDB requests at 3.6s delay (1000/hr max), but BookDB's actual rate limits are much lower:
| Tier |
BookDB Limit |
LM Sends |
| Signed LM, no API key (Tier 2) |
500/day (~21/hr) |
1000/hr |
| Known LM, no signature (Tier 3) |
200/hr |
1000/hr |
| With API key (Tier 1) |
300/hr |
1000/hr |
LM can exceed BookDB's limits by 3-48x depending on the tier, causing 429 responses despite careful pacing.
Fix
Update library_manager/providers/rate_limiter.py — increase min_delay for bookdb from 3.6 to 12.0 (300/hr, matching the most common API key tier). Also add adaptive backoff that reads the actual tier from 429 response headers if available.
Files
library_manager/providers/rate_limiter.py — change min_delaylibrary_manager/providers/bookdb.py — improve 429 handling to extract retry-after header
Cross-reference
BookDB side should also raise tier limits for signed LM clients — that's a separate Skaldleita issue.
Problem
Library Manager paces BookDB requests at 3.6s delay (1000/hr max), but BookDB's actual rate limits are much lower:
LM can exceed BookDB's limits by 3-48x depending on the tier, causing 429 responses despite careful pacing.
Fix
Update
library_manager/providers/rate_limiter.py— increasemin_delayforbookdbfrom3.6to12.0(300/hr, matching the most common API key tier). Also add adaptive backoff that reads the actual tier from 429 response headers if available.Files
library_manager/providers/rate_limiter.py— change min_delaylibrary_manager/providers/bookdb.py— improve 429 handling to extract retry-after headerCross-reference
BookDB side should also raise tier limits for signed LM clients — that's a separate Skaldleita issue.