feat(cache): in-memory tier, size limits, sharding, observability#1122

Open

vibegui wants to merge 19 commits intomainfrom

vibegui/loader-cache-layers

Commits on Mar 18, 2026

feat(cache): add in-memory tier, size limits, hex sharding, and observability

vibegui
and
claude
committed
feat(cache): add comprehensive cache observability metrics

vibegui
and
claude
committed
fix(cache): preserve response status, validate env vars, fix metric ordering

vibegui
and
claude
committed
fix(cache): address code review findings

vibegui
and
claude
committed
fix(cache): address correctness issues in env parsing, shard init, LRU tracking, and tiered headers

vibegui
and
claude
committed
feat(cache): lazy re-index disk entries into LRU on pod restart

vibegui
and
claude
committed
perf(cache): eliminate hot-path allocations and fix correctness bugs

vibegui
and
claude
committed
fix(cache): async mkdir and single-tier cache-tier header

vibegui
and
claude
committed
fix(cache): reject oversized entries before writing to disk

vibegui
and
claude
committed
test(cache): add tests for tiered cache, in-memory cache, and lazy re-index

vibegui
and
claude
committed
fix(cache): evict stale disk entry when oversized write is rejected

vibegui
and
claude
committed
fix(cache): NaN guard, oversized check in L1, memoize LRU open()

vibegui
and
claude
committed
perf(cache): skip body read in L1 put when Content-Length exceeds limit

vibegui
and
claude
committed
feat(cache): extend default stale window to 1h, add STALE_WINDOW_S env var

vibegui
and
claude
committed
fix(cache): bump STALE_TTL_PERIOD default from 30s to 1h

vibegui
and
claude
committed
feat(cache): bot write guard, L1 admission filter, eviction logging

vibegui
and
claude
committed
fix(cache): update inMemoryCache tests for admission filter

vibegui
and
claude
committed
feat(cache): write rate warning and disk fill warning

vibegui
and
claude
committed
fix(cache): separate singleFlight key for bot requests

vibegui
and
claude
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(cache): in-memory tier, size limits, sharding, observability#1122

feat(cache): in-memory tier, size limits, sharding, observability#1122
vibegui wants to merge 19 commits intomainfrom
vibegui/loader-cache-layers

Commits on Mar 18, 2026

feat(cache): add in-memory tier, size limits, hex sharding, and observability

feat(cache): add comprehensive cache observability metrics

fix(cache): preserve response status, validate env vars, fix metric ordering

fix(cache): address code review findings

fix(cache): address correctness issues in env parsing, shard init, LRU tracking, and tiered headers

feat(cache): lazy re-index disk entries into LRU on pod restart

perf(cache): eliminate hot-path allocations and fix correctness bugs

fix(cache): async mkdir and single-tier cache-tier header

fix(cache): reject oversized entries before writing to disk

test(cache): add tests for tiered cache, in-memory cache, and lazy re-index

fix(cache): evict stale disk entry when oversized write is rejected

fix(cache): NaN guard, oversized check in L1, memoize LRU open()

perf(cache): skip body read in L1 put when Content-Length exceeds limit

feat(cache): extend default stale window to 1h, add STALE_WINDOW_S env var

fix(cache): bump STALE_TTL_PERIOD default from 30s to 1h

feat(cache): bot write guard, L1 admission filter, eviction logging

fix(cache): update inMemoryCache tests for admission filter

feat(cache): write rate warning and disk fill warning

fix(cache): separate singleFlight key for bot requests