@aikirun/worker 0.9.0 → 0.9.1

package/README.md

@@ -1,7 +1,6 @@
 # @aikirun/worker
 
-Worker SDK for Aiki durable execution platform - execute workflows and tasks with durable state management and automatic
-recovery.
+Worker SDK for Aiki durable execution platform.
 
 ## Installation
 
@@ -11,138 +10,48 @@ npm install @aikirun/worker
 
 ## Quick Start
 
-### Create and Spawn a Worker
-
 ```typescript
 import { worker } from "@aikirun/worker";
 import { client } from "@aikirun/client";
-import { onboardingWorkflowV1 } from "./workflows.ts";
-
-// Define worker
-const aikiWorker = worker({
-	name: "worker-1",
-	workflows: [onboardingWorkflowV1],
-	subscriber: { type: "redis" },
-	opts: {
-		maxConcurrentWorkflowRuns: 10,
-	},
-});
+import { orderWorkflowV1 } from "./workflows.ts";
 
-// Initialize client
 const aikiClient = await client({
 	url: "http://localhost:9876",
 	redis: { host: "localhost", port: 6379 },
 });
 
-// Spawn worker
-const handle = await aikiWorker.spawn(aiki);
-```
-
-### Graceful Shutdown
+const aikiWorker = worker({
+	name: "order-worker",
+	workflows: [orderWorkflowV1],
+});
 
-```typescript
-import process from "node:process";
+const handle = await aikiWorker.spawn(aikiClient);
 
-const shutdown = async () => {
+// Graceful shutdown
+process.on("SIGTERM", async () => {
 	await handle.stop();
-	await aiki.close();
+	await aikiClient.close();
 	process.exit(0);
-};
-
-process.on("SIGINT", shutdown);
-process.on("SIGTERM", shutdown);
-```
-
-## Features
-
-- **Durable Execution** - Automatically resume from failures without losing progress
-- **Horizontal Scaling** - Multiple workers process workflows in parallel
-- **State Management** - Persist execution state at each step
-- **Automatic Recovery** - Detect stuck workflows and retry automatically
-- **Polling Strategies** - Adaptive polling with configurable backoff
-- **Graceful Shutdown** - Clean worker termination with in-flight workflow handling
-
-## Horizontal Scaling
-
-Scale workers by creating separate definitions to isolate workflows or shard by key:
-
-```typescript
-// Separate workers by workflow type
-const orderWorker = worker({ name: "orders", workflows: [orderWorkflowV1] });
-const emailWorker = worker({ name: "emails", workflows: [emailWorkflowV1] });
-
-await orderWorker.spawn(client);
-await emailWorker.spawn(client);
-```
-
-```typescript
-// Shard workers by key (reuse base definition with different shards)
-const orderWorker = worker({ name: "order-processor", workflows: [orderWorkflowV1] });
-
-await orderWorker.with().opt("shards", ["us-east", "us-west"]).spawn(client);
-await orderWorker.with().opt("shards", ["eu-west"]).spawn(client);
-```
-
-## Worker Configuration
-
-### Params (required for worker identity)
-
-```typescript
-interface WorkerParams {
-	name: string; // Unique worker name
-	workflows: WorkflowVersion[]; // Workflow versions to execute
-	subscriber?: SubscriberStrategy; // Message subscriber (default: redis)
-}
-```
-
-### Options (via `opts` param or `with()` builder)
-
-```typescript
-interface WorkerOptions {
-	maxConcurrentWorkflowRuns?: number; // Concurrency limit (default: 1)
-	workflowRun?: {
-		heartbeatIntervalMs?: number; // Heartbeat interval (default: 30s)
-	};
-	gracefulShutdownTimeoutMs?: number; // Shutdown timeout (default: 5s)
-	shards?: string[]; // Optional shards for distributed work
-}
-```
-
-## Workflow Registration
-
-Workers receive workflow versions through the `workflows` param:
-
-```typescript
-const aikiWorker = worker({
-	name: "worker-1",
-	workflows: [workflowV1, workflowV2, anotherWorkflowV1],
-	subscriber: { type: "redis" },
 });
 ```
 
-The worker automatically discovers and executes the registered workflow versions.
-
-## State Persistence
+## Features
 
-Workers store execution state at each step:
+- **Horizontal Scaling** - Run multiple workers to share workload
+- **Automatic Recovery** - Resume from failures without losing progress
+- **Heartbeat Monitoring** - Detect and recover stuck workflows
+- **Graceful Shutdown** - Complete active work before stopping
+- **Sharding** - Route workflows to specific workers
 
-- Task completion status
-- Sleep/wait checkpoints
-- Event acknowledgments
-- Child workflow results
+## Documentation
 
-This allows workflows to resume from the exact point of failure.
+For comprehensive documentation including scaling strategies, configuration options, and how workers operate, see the [Workers Guide](https://aiki.run/docs/core-concepts/workers).
 
 ## Related Packages
 
-- [@aikirun/client](https://www.npmjs.com/package/@aikirun/client) - Start workflows
+- [@aikirun/client](https://www.npmjs.com/package/@aikirun/client) - Connect to Aiki server
 - [@aikirun/workflow](https://www.npmjs.com/package/@aikirun/workflow) - Define workflows
 - [@aikirun/task](https://www.npmjs.com/package/@aikirun/task) - Define tasks
-- [@aikirun/types](https://www.npmjs.com/package/@aikirun/types) - Type definitions
-
-## Changelog
-
-See the [CHANGELOG](https://github.com/aikirun/aiki/blob/main/CHANGELOG.md) for version history.
 
 ## License
 

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@aikirun/worker",
-	"version": "0.9.0",
+	"version": "0.9.1",
 	"description": "Worker SDK for Aiki - execute workflows and tasks with durable state management and automatic recovery",
 	"type": "module",
 	"main": "./dist/index.js",
@@ -18,9 +18,9 @@
 		"build": "tsup"
 	},
 	"dependencies": {
-		"@aikirun/types": "0.9.0",
-		"@aikirun/client": "0.9.0",
-		"@aikirun/workflow": "0.9.0"
+		"@aikirun/types": "0.9.1",
+		"@aikirun/client": "0.9.1",
+		"@aikirun/workflow": "0.9.1"
 	},
 	"publishConfig": {
 		"access": "public"