Durable Objects are getting faster every week, with a ton of work going into optimisations ๐
This was my last change before PTO, a ~45% reduction in p50/p99 latency on cache miss for Durable Objects that have been relocated
I'm new to the Durable Objects team, and I've been seriously amazed by how locked in everyone is on making DOs and Workers the best platform to build on.
Two big wins delivered from the team - together, these help your app & agent get to work sooner.
Both of these improvements are live, with no code changes needed.
fewer cold starts:
https://t.co/QGrHirr1jB
improved relocation request routing:
https://t.co/5Fy2JEC2h5
@BertosonHunter@mynameistito@Cloudflare The TLDR is lots of things - running agents on them right now is a big one especially if you've got lots of horizontal scaling needs or just want all the benefits of isolates (cheap) with persisted state! I use them for my personal agents
@BertosonHunter@mynameistito@Cloudflare I've heard this a few times. I've got this draft doc update I just put up a PR for. Feedback is welcome whether or not it helps anything click? lmk.
https://t.co/uAb6l6dRHI
Our runtime now retries fetch() and RPC calls when a retryable error prevents them from reaching a Durable Object, so you donโt have to worry about that retry logic yourself anymore.
For methods that are fine to run more than once, you can use retryable() to let them run again on a retry.
https://t.co/zAa8P1GuyJ
SurrealDB 4 and @SurrealDB Earth are out today.
When we started SurrealDB, I wanted one database that could hold every shape of data and query it all together.
Now every agent, user and app can have a database of its own.
๐ด rest easy - Durable Objects can now wait until they hibernate to apply code updates, giving long-running agent turns time to finish even if you're shipping PRs by the second.
https://t.co/sjIlhw7W4f