1.7 KiB
1.7 KiB
id, title, status, assignee, created_date, labels, dependencies, references, priority
| id | title | status | assignee | created_date | labels | dependencies | references | priority | |
|---|---|---|---|---|---|---|---|---|---|
| ML-16 | Inconsistent HTTP timeout configuration across API clients | To Do | 2026-04-20 08:50 |
|
low |
Description
GitHub: created 2026-04-05 · updated 2026-04-09 · closed 2026-04-08 · not planned
Summary
HTTP timeout configuration varies significantly across API integrations with no documented rationale for the differences.
Evidence
- Last.fm: Custom
pool_timeout: 10_000,receive_timeout: 2500,connect_timeout: 2500 - OpenAI GPT:
receive_timeout: 10_000,connect_timeout: 2_500 - OpenAI Chat stream:
receive_timeout: 60_000,connect_timeout: 5_000 - MusicBrainz, Discogs, Wikipedia, Brave Search: Default timeouts only (no custom configuration)
Why This Matters
- No consistency in how timeouts are configured
- Unclear whether defaults are appropriate for each API's characteristics
- Last.fm applied custom timeouts based on observed issues; other APIs may have similar unaddressed problems
Suggested Fix
- Audit each API's typical response times
- Set explicit, documented timeout values for all API clients
- Consider making timeouts configurable via each API's Config module
Acceptance Criteria
- All API clients have explicit, intentional timeout configuration
- Timeout choices are documented (e.g., in Config module comments)
- #1 All API clients have explicit, intentional timeout configuration
- #2 Timeout choices are documented (e.g., in Config module comments)