Dapr

Looking for a hands-on way to learn Dapr Workflow in .NET with Aspire? The new free Dapr University track runs in a browser, so it's easy to run without installing anything locally. Start learning now: https://lnkd.in/eFKTD_T7

Missed the Dapr Community Call this week? You can now watch the recording on YouTube: https://lnkd.in/epFSTHrj - Dapr case study EVA Systems - Dapr Workflow Access Policy - Community show and tell Join the Dapr Discord to continue the conversation: https://lnkd.in/eFYKm2wi

🚀 Dapr 1.18 is here — and it's all about Workflows! This release puts Dapr Workflows front and center, hardening them for security, durability, and scale. Whether you're running long-running business processes, human-in-the-loop approvals, or AI agent orchestration, 1.18 gives your workflows enterprise-grade guarantees out of the box. Here's what's new for Workflows 👇 📦 Workflow History Context Propagation Workflows can now pass slices of their history down to children. Child workflows read the upstream history context through a typed API — no more threading state through inputs or reaching into a state store. 🛡️ Workflow History Tamper Detection Workflow history can now be cryptographically signed and verified on every state load. Each step is signed under the sidecar's mTLS SPIFFE identity and chained to the previous one — so tampering is caught the moment state is read. Protection extends across app boundaries too, with child workflow and activity completion attestation. 🔐 WorkflowAccessPolicy A new CRD that controls which app IDs can invoke which workflows and activities. A pure allow-list with per-operation rules and glob matching — purpose-built for shared and multi-tenant clusters where unrestricted cross-app calls are a risk. 🎚️ Workflow Concurrency Limits Cap how many workflows or activities run at once across the entire cluster, with global and per-name limits. The scheduler queues triggers and dispatches them as slots free up — backpressure done right. 🙋 Human-in-the-Loop, Now Observable Indefinite WaitForExternalEvent waits now materialize a synthetic timer tagged with the awaited event name — so you can finally see what an instance is blocked on. Approval gates and human-in-the-loop patterns just got operable. 🪂 Graceful Stall on Oversized Payloads Oversized workflow payloads used to tear down the whole stream and cancel every pending workflow. Now they're gracefully stalled in a recoverable state — the rest of the stream keeps running, and no instance is lost. And beyond Workflows: the Jobs API graduates to Stable ✅, hot-reloading goes GA (on by default) 🔄, native Kubernetes sidecar container support, IPv6/dual-stack host resolution, and pub/sub now drains in-flight messages on graceful shutdown. 💜 Thank you to our contributors and our community! Dapr is built by an incredible community, and 1.18 is no exception. Huge thanks to everyone who contributed code, docs, reviews, and feedback this cycle. Read the full release notes in the blog post: https://lnkd.in/evaQwZGm

Join the Dapr Community Call on June 10th. Continue the discussion on the Dapr Discord: https://lnkd.in/eFYKm2wi

Join the Dapr Community Call today (June 10th)! - Dapr case study EVA Systems - Workflow Access Policies - .NET SDK udpate Join at: https://lnkd.in/e9FgAQRZ

Join the Dapr Community Call tomorrow (June 10th)! - Dapr case study EVA Systems - Workflow Access Policies - .NET SDK udpate Join at: https://lnkd.in/e9FgAQRZ

Join the Dapr Community Call on June 10th! - Dapr case study EVA Systems - Workflow Access Policies - .NET SDK udpate Join at: https://lnkd.in/e9FgAQRZ

Looking for a hands-on way to learn Dapr Workflow in .NET with Aspire? The new free Dapr University track runs in a browser sandbox, so it's easy to run without installing anything locally. Start learning now: https://lnkd.in/eFKTD_T7

Looking for a hands-on way to learn Dapr Workflow in .NET with Aspire? The new free Dapr University track runs in a browser sandbox, so it's easy to run without installing anything locally. Start learning now: https://lnkd.in/eFKTD_T7

Today we’re announcing a durable workflow integration between Dapr Workflows and Claude Managed Agents (Link in the comments 👇) 🚀 This integration allows developers to merge advanced agent reasoning with a durable orchestration engine tailored for distributed systems, featuring: * Automatic recovery after crashes or restarts * Multi-instance workflow coordination * Durable timers and external events * Replayable execution history * Long-running workflows that endure infrastructure changes * Resilient orchestration across agents, MCP servers, tools, and services This development underscores a significant gap in the agent ecosystem: persisting state alone is insufficient. Many frameworks, including Claude Managed Agents, only persist conversation history or checkpoints and give you basic APIs to resume it, putting the failure detection and recovery burden on you. But when an agent process crashes at 2 AM while orchestrating multiple MCP servers, coordinating APIs, and operating across various Kubernetes instances, what happens next? * Does execution resume automatically after an infrastructure failure? * Can another worker seamlessly continue execution? * Do timers survive process termination? * Can distributed workflow instances coordinate safely? This integration clearly shows the distinction between merely storing state and having a production-grade durable workflow runtime. Dapr now supports first-class integrations with: * OpenAI Agents * Google ADK * Strands Agents * Microsoft Agent Framework * Pydantic AI * Claude Managed Agents * HolmesGPT SRE agents * LangGraph * LangChain Deep Agents * CrewAI Our goal is to make Dapr the default choice for durable AI workloads, and broad ecosystem integrations are a critical piece. More integrations coming soon 😎