Most AI agent tools let you watch your agents fail ๐ฟ
Restate lets you do something about it ๐ฎ
โธ๏ธ Pause during outages
โ Kill stuck retry loops
๐ Restart from a specific step, keep the work done
๐ฌ Update in-flight agent calls
Your agents deserve a remote control: www.restate.dev/blog/a-remot...
Posts by Till Rohrmann
@restate.dev v1.6.1 + v1.6.2 just dropped ๐
โ ๏ธ fixed leadership thrashing that could destabilize clusters
๐ง slashed per-invocation memory (gigabytes saved)
๐ fixed lazy state keys stuck in retry loops
๐ฆ weekly Docker security refreshes
details โ github.com/restatedev/r...
Restate + Vercel AI SDK =
โ
Recover from failures without losing progress
โ
Built-in state management
โ
Trace every decision and action
โ
From simple agents to complex multi-agent systems
โ
Resilient human approvals, timeouts, and rollback
www.restate.dev/blog/buildin...
Restate + Vercel AI SDK = enterprise-grade durability for your AI agents with just a few lines of code.
Survive crashes, timeouts, and restarts - even for multi-turn conversations, human approval and long-running research agents.
@vercel.com @sdk.vercel.ai
AWS's recent us-east-1 outage hit ChatGPT, Snapchat, and Perplexity. In 2021: Netflix, Disney+, Robinhood.
The problem? No true geo-replication.
We show how @restate.dev makes it simple with zero downtime during a simulated region failure:
www.restate.dev/blog/geo-rep...
If yesterday's AWS outage made you rethink your infrastructure, this blog post shows exactly how to build apps that stay up when the cloud goes down.
๐ Full blog post: www.restate.dev/blog/geo-rep...
AWSโs us-east-1 outage took down ChatGPT, Snapchat & Perplexity.
The same happened in 2021: Netflix, Disney+, and Robinhood all went dark.
The reason? Often a lack of true geo-replication.
Here's how you can run apps with zero downtime during a region failure using Restate ๐งต
Webhooks that just work with @restate.dev
โ
Event persistance - let Restate acknowledge and persist the event
โ
Reliable processing - leverage Durable Execution to run until completion
โ
Idempotency - use Idempotency-Key to deduplicate
Check it out! docs.restate.dev/guides/durab...
Restate Cloud has all you need to develop transactional, resilient, secure applications.
Connect services running on serverless compute platforms like @cloudflare.social, @vercel.com, @deno.land, or AWS Lambda. Or on your own infra on AWS, GCP, or Kubernetes.
www.restate.dev/blog/announc...
๐ Exciting news! ๐ฅ๐ฒ๐๐๐ฎ๐๐ฒ ๐๐น๐ผ๐๐ฑ is now publicly available! Build workflows and AI agents with durable execution, observability, and production-grade availability without running any stateful infra yourself.
๐ Yesterday, we announced that ๐ฅ๐ฒ๐๐๐ฎ๐๐ฒ ๐๐น๐ผ๐๐ฑ is now publicly available. Today, weโre announcing ๐ฅ๐ฒ๐๐๐ฎ๐๐ฒ ๐ญ.๐ฑ!
This release brings key features for building, controlling and observing your durable agents and workflows.
A lot is brewing at Restate! Hereโs the first in a series of rollouts coming over the next few weeks โจ Stay tuned!
๐ Restate v1.4.3 is here!
โจ New features:
๐ง Shell completions for CLI tools
โต Enhanced Helm charts
๐ OTLP+HTTP tracing support
Seamless upgrade, no migration needed! ๐
๐ฆ Download: github.com/restatedev/r...
๐ Get started: docs.restate.dev/get_started/...
@stephanewen.bsky.social will give us a demo of Restate + AI Agents and how you can get started!
Register here: lu.ma/372aiucf
Join the community meeting next week to learn how combining Restate with your agent SDK (Vercel AI, OpenAI, ...) gives your agents a set of powerful features.
๐ Restate 1.4 just dropped! 10x faster failure detection, automatic workload rebalancing, and 2.5x quicker startup. Your distributed applications just got more resilient with less downtime and better resource utilization. Build more, worry less. #Restate #Resilience restate.dev/blog/announc...
Watch the community meeting on Restate 1.3 here:
Prefer AWS Lambda or another FaaS provider? Have a look at this previous blog post we wrote:
restate.dev/blog/we-repl...
#serverless #AWSLambda
Restate lets you write resilient sagas in clean, readable code that handles failures gracefully.
Check out the updated sagas guide: docs.restate.dev/guides/sagas/
#DistributedSystems #Microservices #Sagas
Restate auto-retries temporary issues like network glitches. But for business logic failures ("Hotel fully booked"), retrying won't help. Instead, Restate helps you gracefully undo previous steps using sagas - each action has a matching "undo" step that runs when needed.
Hey everyone, Iโll be in New York on June 24/25. If you want to talk about @restate.dev, durable execution, challenges in how to productionize agentic workflows or just grab a coffee, then let me know. Happy to meet you in NY ๐
Read the full guide here ๐
restate.dev/blog/buildin...
#serverless #Knative #kubernetes #devinfra
Serverless is great, until you need to manage state. This is where Restate comes in!
Learn how to combine Knative with Restate to build serverless applications on Kubernetes that are not only scalable, but also stateful and resilient. ๐งต
Exploring @restate.dev as a replacement for some long-running, complex, asynchronous orchestrators currently implemented in Step Functions. This 2023 post resonates hard. Lambda chain -> Step Functions is an increasingly common migration but comes up short vs RPC/ECS restate.dev/blog/we-repl...
There is a new Restate 1.3.1 release ๐
It contains various fixes and quality of life improvements:
* Improved Admin API OpenAPI spec
* Improved file corruption protection
* Protection against accidentally starting Restate from a wrong working dir
Get it while it's ๐ฅ
github.com/restatedev/r...
๐ Big news! @restate.dev 1.3 is here โ and it makes handling concurrency a breeze. Complex flows? Restate journals them, making everything replayable and fault-tolerant so you can sleep easy.
Dive in ๐ restate.dev/blog/announc... to see whatโs new!
Restate service handler written in Kotlin, showing how to implement a scatter/gather use case with few lines of code in a resilient way.
Coming up in @restate.dev 1.3: Implement scatter/gather scenarios with ease, while keeping your application resilient to failures! Those sub tasks can fail, the orchestrator itself can fail, and yet this code will run to completion thanks to Restate "run to completion" semantics.