Elie Habib e48162e9c0 fix(rate-limit): add per-endpoint rate limits for summarize endpoints (#1161) ...

Cache lookups (summarize-article-cache) shared the global 600/60s IP
rate limit with expensive LLM calls (summarize-article), causing users
to burn rate limit budget on cheap Redis reads.

- Add ENDPOINT_RATE_POLICIES map with per-endpoint limits
- summarize-article POST: 20 req/60s (protect LLM quota)
- summarize-article-cache GET: 3000 req/60s (generous for browsing)
- Two-phase gateway check: endpoint-specific first, global fallback
- Endpoints with specific policy skip global limiter entirely
- Path normalization prevents trailing-slash bypass

2026-03-07 08:51:45 +04:00

..

acled.ts

refactor(server): consolidate declare const process into shared env.d.ts (#752)

2026-03-02 14:01:41 +04:00

cache-keys.ts

fix(economic): seed all WB indicators on Railway, never call WB API from frontend (#1159)

2026-03-07 08:00:28 +04:00

constants.ts

refactor: deduplicate clampInt into server/_shared/constants

2026-03-03 01:11:44 +04:00

hash.ts

fix(security): harden cache keys against injection and hash collision (WM-2026-001/002) (#1103)

2026-03-06 10:33:41 +04:00

rate-limit.ts

fix(rate-limit): add per-endpoint rate limits for summarize endpoints (#1161)

2026-03-07 08:51:45 +04:00

redis.ts

fix(health): write seed-meta on cache hits (throttled 5min) (#1138)

2026-03-06 23:49:51 +04:00

response-headers.ts

feat(gateway): complete edge cache tier coverage + degraded-response policy (#484)

2026-02-28 00:34:58 +04:00