A run is only as reliable as the work inside it, including the sandbox that tests the code.
Today's sandboxes live outside the durable execution chain. They can't see the run that called them, which means:
1) you're billed through its waits
2) durability is left to your code
3x concurrency when you need it most ↗️
Spikes are pretty unpredictable, but not a bad thing!
It usually just means your product is getting used more than you expected.
Bursty concurrency in Inngest raises your limit by 𝟯𝘅 as soon as a spike begins, so you (and your users)
We’re at The AI Conference 👋
Come find the Inngest crew at booth #136 to talk unbreakable agents, long-running workflows, and why reliability shouldn’t need its own architecture.
And yes, the cowboy hat made it to San Francisco. 🤠
See you there.
We went to Portland. We asked 3 teams to build something that keeps running while they do something else. Things got weird 😅
Our episode of the Web Dev Challenge is now live! Huge thanks to the whole crew, and our hilarious contestants 😆