Durable Objects aren't just durable, the...
I just read: https://blog.cloudflare.com/how-we-built-cloudflare-queues/
Also the part "Workers KV acts as a fast metadata store: our Worker can quickly look up the shard map to perform load balancing across shards"
How can they use KV for that, as KV has a 60 second delay to update after writing? Won't the entire infrastructure have a 60 second delay to understand what the state is and how would the be able to update that state?
The Cloudflare Blog
Durable Objects aren't just durable, they're fast: a 10x speedup fo...
Learn how we built Cloudflare Queues using our own Developer Platform and how it evolved to a geographically-distributed, horizontally-scalable architecture built on Durable Objects. Our new architecture supports over 10x more throughput and over 3x lower latency compared to the previous version.
4 Replies
This is one of the limits (https://developers.cloudflare.com/kv/platform/limits/) that we can adjust on a per-customer basis for very high throughput workloads (>1000 RPS)
Cloudflare Docs
Limits | Cloudflare Workers KV
got it, but you will still have the issue with reads right? or did they solve getting faster consistency in a other way or was that not needed?
There is still some degree of eventual consistency in terms of propagation across KV's central storage backends, but this is tolerated within the Queues design because sharding occurs while there is still room to grow within existing Durable Objects
Thanks for explaining. Really love these blogs!!!