decentraland / comms-gatekeeper / 25082324459

84%

main: 85%

Build Type

Pull #258

github

Committed by Jarvis

Commit Message

fix: reuse LRU cache instances across requests in lands and worlds adapters

cachedFetch.cache() was called inside function bodies, creating a new
empty LRU cache on every request — so every call was a guaranteed cache
miss and a fresh HTTP fetch. This caused comms-gatekeeper to generate
180K+ uncached requests/hr to the Catalyst for /parcels/{x}/{y}/operators
and other endpoints.

Move cache instantiation to component initialization scope, matching the
existing correct pattern in places.ts and denylist.ts. With the default
5-minute TTL, this reduces repeated lookups for the same URL from N
HTTP calls to 1 per TTL window.

Affected call sites (5 total):
- lands.ts: getLandPermissions, getLandOperators
- worlds.ts: fetchWorldActionPermissions, fetchWorldSceneEntityMetadataById, hasWorldOwnerPermission

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Pull Request Pull Request #258: fix: reuse LRU cache instances across requests

Coverage Stats

1018 of 1331 branches covered (76.48%)

Branch coverage included in aggregate %.

9 of 10 new or added lines in 2 files covered. (90.0%)

5 existing lines in 1 file now uncovered.

2852 of 3284 relevant lines covered (86.85%)

42.23 hits per line