AmanTahiliani 7f0c338a57 Add OpenF1 request pacing, 429 retry, and hub completeness guard ...

Concurrent fan-outs could burst past the free-tier rate limit; callers
swallowed the 429s as missing data, and the championship hub cached the
resulting partial season for its full TTL (observed live: 8 of 22
rounds, zero wins/podiums/form).

- requestPacer spaces live requests (350ms anonymous, 100ms with key);
  cache hits never wait
- get() retries 429s up to 3 times honouring Retry-After (capped 10s)
- hub tracks per-meeting fetch failures and caches incomplete
  aggregates for only 2 minutes instead of 15min/24h

Verified against the live API: 2026 hub now reports all 22 rounds with
wins/podiums/form populated.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

2026-07-03 00:47:52 -04:00

..

cache.go

feat(backend): implement session coverage, resumable ingestion, deep year backfill, backoff rate-limit, and dynamic cache TTLs

2026-05-25 13:52:43 -04:00

client.go

Add OpenF1 request pacing, 429 retry, and hub completeness guard

2026-07-03 00:47:52 -04:00

openf1_test.go

Open F1 API calls

2026-03-03 00:20:09 -05:00

openf1.go

Add OpenF1 request pacing, 429 retry, and hub completeness guard

2026-07-03 00:47:52 -04:00

pacing_test.go

Add OpenF1 request pacing, 429 retry, and hub completeness guard

2026-07-03 00:47:52 -04:00

session_selection_test.go

Add race control visuals and driver names in laps view

2026-05-25 15:55:36 -04:00