[Feature]: Native AWS Lambda Powertools for TypeScript integration

[abhishekc-timedotmoney]

Feature Request

Native integration with AWS Lambda Powertools for TypeScript Logger, similar to how the Python Durable SDK supports Powertools.

Problem

When using Powertools Logger as a customLogger via context.configureLogger(), the logs get double-wrapped in CloudWatch because:

1. Durable functions require LogFormat: JSON in the Lambda config
2. Lambda's LogFormat: JSON patches the global console object to wrap output in {"timestamp", "level", "requestId", "message"}
3. Powertools Logger uses the global console.info() to output its own structured JSON
4. Lambda catches that and stringifies Powertools' JSON into the message field

Result — double-wrapped JSON:

{
  "timestamp": "2026-03-09T13:52:31.481Z",
  "level": "INFO",
  "requestId": "fa3b7f07-0be3-4380-9a53-fb080779fcf4",
  "message": "{\n    \"level\": \"INFO\",\n    \"message\": \"workflow-started\",\n    \"timestamp\": \"2026-03-09T19:22:31.039+05:30\",\n    \"service\": \"lambda-user-account-deletion\",\n    \"sampling_rate\": 0,\n    \"xray_trace_id\": \"1-69aed09e-5a04f2fc1f7bee1748385655\",\n    \"tm_user_id\": \"0199c820-c4bd-778a-9197-580108aea912\"\n}"
}

Workaround

Override the Powertools Logger's internal console with a direct process.stdout writer (same technique the Durable SDK's own DefaultLogger uses and Powertools itself uses):

import { Console } from 'node:console';
import { Logger } from '@aws-lambda-powertools/logger';

const logger = new Logger({ serviceName: 'my-service' });

// Bypass Lambda's LogFormat:JSON console patching to prevent double-wrapped JSON logs
(logger as unknown as { console: Console }).console = new Console({
  stdout: process.stdout,
  stderr: process.stderr,
});

export { logger };

Then pass it as customLogger:

const handler = async (event: MyEvent, context: DurableContext<Logger>) => {
  context.configureLogger({ customLogger: logger });
  context.logger.info({ message: 'workflow-started' });
};

export const myFunction = withDurableExecution(handler);

Result — clean Powertools output:

{
    "level": "INFO",
    "message": "workflow-started",
    "timestamp": "2026-03-09T19:37:51.268+05:30",
    "service": "my-service",
    "sampling_rate": 0,
    "xray_trace_id": "1-69aed434-506caa866976504e1a0ec55f"
}

Requested Feature

1. First-class Powertools integration — The Python Durable SDK has native Powertools support where context.configure_logger(powertools_logger) automatically injects durable execution metadata (executionArn, operationId, attempt, requestId) into every Powertools log entry via append_keys(). The JS SDK should have parity.

2. Handle the console patching conflict — When Powertools is detected as the customLogger, the SDK should automatically handle the LogFormat: JSON console patching conflict so users don't need the workaround above.

3. Inject durable metadata into Powertools structured logs — Use Powertools' appendKeys() method to inject durable execution context (like executionArn, operationId) into every log entry automatically, rather than just delegating log calls.

Environment

• SDK Version: latest
• Node.js: 24.x
• Powertools: @aws-lambda-powertools/logger v2
• Runtime: Durable Function (nodejs:24.DurableFunction.v10)

References

• Python SDK Powertools integration: https://github.com/aws/aws-durable-execution-sdk-python/blob/main/docs/core/logger.md
• Related issue Populate structured log data in custom object #305: Populate structured log data in custom object
• Related issue [Bug]: Structured logger adds unneccessary \n to logging payloads #322: Structured logger adds unnecessary \n to logging payloads
• Powertools console bypass: https://github.com/aws-powertools/powertools-lambda-typescript/blob/main/packages/logger/src/Logger.ts#L1102-L1105