What is the core rule for separating Workflows and Activities?

Workflows should only contain orchestration logic and decision-making and must be deterministic. Activities handle all external interactions like API calls, database writes, and network requests.

How does this skill handle distributed transaction failures?

It utilizes the Saga pattern with compensating actions, where each successful step registers a corresponding undo action that runs in reverse order (LIFO) if a failure occurs later in the process.

Why is determinism critical in workflow orchestration?

Orchestrators like Temporal use event replaying to restore state after a crash. If the code is non-deterministic (e.g., using random numbers or system time), the replay would diverge from the original execution, causing failures.

When should I avoid using workflow orchestration?

Orchestration is overkill for simple CRUD operations, stateless request/response APIs, or high-volume real-time streaming data that is better handled by tools like Kafka.

Workflow Orchestration Patterns

Name: Workflow Orchestration Patterns
Author: HermeticOrmus

byHermeticOrmus

API開発

Design and implement resilient distributed systems using Temporal workflow patterns, saga logic, and state management best practices.

概要

This skill provides expert architectural guidance for building durable, distributed systems using the Temporal orchestration engine. It helps developers navigate the critical distinction between deterministic orchestration logic (Workflows) and non-deterministic external interactions (Activities), while providing implementation patterns for complex scenarios like the Saga pattern for distributed transactions and human-in-the-loop callbacks. By following these patterns, you can ensure that your long-running processes are automatically recoverable, scalable, and maintainable across microservice boundaries.

主な機能

Design patterns for separating orchestration logic from external side effects
Strategies for parallel execution using fan-out/fan-in and child workflows
Best practices for maintaining workflow determinism and state persistence
Implementation guides for the Saga pattern with compensating actions for rollbacks
Advanced failure recovery configuration including retry policies and activity heartbeats
0 GitHub stars

ユースケース

Orchestrating multi-step distributed transactions like payment processing or order fulfillment
Implementing human-in-the-loop workflows requiring manual approvals and escalations
Managing long-running entity lifecycles such as user account states or inventory tracking

概要

主な機能

Design patterns for separating orchestration logic from external side effects
Strategies for parallel execution using fan-out/fan-in and child workflows
Best practices for maintaining workflow determinism and state persistence
Implementation guides for the Saga pattern with compensating actions for rollbacks
Advanced failure recovery configuration including retry policies and activity heartbeats
0 GitHub stars

ユースケース

Orchestrating multi-step distributed transactions like payment processing or order fulfillment
Implementing human-in-the-loop workflows requiring manual approvals and escalations
Managing long-running entity lifecycles such as user account states or inventory tracking