durable-execution

Implement crash recovery for long-running workflows with durable checkpoints and replay.

6|Updated Jan 8, 2026
One-click install
npx skills add https://github.com/resonatehq/resonate-skills --skill durable-execution
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: durable-execution
Source: https://github.com/resonatehq/resonate-skills/tree/main/durable-execution
Command: npx skills add https://github.com/resonatehq/resonate-skills --skill durable-execution

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve?

Durable execution ensures that when a process crashes, it restarts from the last checkpoint rather than from the beginning, enabling reliable long-running workflows.

Core Features & Use Cases

  • Durable checkpoints: each step can be replayed from the last completed point using deterministic IDs.
  • Cross-service coordination: orchestrate and coordinate steps across multiple services or worker groups.
  • Durable timers: suspensions and delays survive crashes, ensuring time-based workflows continue when faulted.
  • Human-in-the-loop: suspend workflows awaiting external input or approval without holding resources.
  • Outbox pattern: safely handle side effects (emails, webhooks) without duplicating actions on replay.
  • Saga, fan-out/fan-in, and entity lifecycle patterns: robust patterns for complex multi-step processes.
  • Quick-start templates: starter templates for gateway, workers, and common durability patterns.

Quick Start

Register a generator-based workflow with the Resonate SDK, start the server, and invoke the workflow to observe durable replay across crashes.

Quick Start

Read the pre-built templates and spawn a worker to run a sample durable workflow, observing replay and recovery behavior.

Frequently Asked Questions about durable-execution

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make long-running workflows crash-safe so they restart from the last checkpoint?

Durable execution enables crash recovery for long-running workflows by replaying steps from the last checkpoint using deterministic IDs, ensuring processes restart safely without losing progress.

What is the best way to orchestrate cross-service coordination for multi-step processes?

Cross-service coordination orchestrates steps across multiple services or worker groups using durable execution, ensuring reliable execution and crash recovery for distributed multi-step workflows.

How do I implement human-in-the-loop gates without holding resources during external approval?

Human-in-the-loop gates suspend workflows awaiting external input or approval without holding resources, leveraging durable timers and replay semantics to resume safely after input arrives.

Can I prevent duplicate side effects like emails and webhooks when replaying faulted workflows?

The outbox pattern safely handles side effects like emails and webhooks without duplicating actions during replay, ensuring idempotent side effects across crash recovery cycles.

Does durable execution support generator-based workflows with deterministic child IDs?

Durable execution enforces generator-based workflows with deterministic child IDs and preload replay caches, enabling fast replay and reliable step reconstruction across crash incidents.

What patterns are available for complex multi-step processes like fan-out fan-in and sagas?

Durable execution supports saga, fan-out/fan-in, and entity lifecycle patterns for complex multi-step processes, providing robust templates for reliable distributed workflow coordination.