* Try different approach
* Revalidate before cache expires
* TS
* Only replicate to local backends
* Ensure revalidate is done only once for the same key
* Improve perf of getCacheKey
* Simplify type
* Use better method and comment
* Use it 7 days only
* Add unit tests for "race"
* Reduce to 80s
* Implement a Cloudflare KV cache
* Format
* Lint
* Fix tagging of entries in KV
* No longer use HTTP cache tags
* Simplify tags to only keep 2
* Implement a blockFallback logic
* Improve the block fallback logic
* More tests
* Start introducing a always optional signal on all cache ops
* Correctly pass signal end to end
* Fix one more case and add tests
* Fix it
* Add test for error
* Fix error handling
* Simplify even more
* Improve logs
* Improve logs / measurements
* Fix timing
* Fix read cache duration and add minor tests
* Log redis time
* Change replication logic
* Fix redis errors
* Log more
* More logs and test memory first
* Improve tracing for cache backends
* Ignore all dependencies
* Log the key in the cache low level traces
* Disable browser caching
* Improve cache read
* Fix
* Try using it for lookup
* Read before increment in redis
* Don't use multi
* Limit cache to 2min in cloudflare
* Cache in cloudflare but not in the browser
Johan Preynat