Hi everyone, I’m trying to set up middy with SST and I’m running into some problems. If I forego using middy, SST doesn’t have an issue. However, after adding the middy.js middleware in for the lambda handlers, the lambdas seem to not be getting the request and I’m seeing this error on the console. Does anyone have experience with this kind of problem? (My best guess is this might be failing somewhere in a request/response logger middleware, but it doesn’t specify anything)

Uncaught Exception 	{"errorType":"Error","errorMessage":"The worker thread exited","stack":["Error: The worker thread exited","    at Worker.onWorkerExit (/Users/Michael/workspace/theater-profiles/.sst/artifacts/mrooney-theatrum-profiles--core-stack-CreateProfileLambda/src/lambda/create-profile/handler.js:1785:36)","    at Worker.emit (node:events:394:28)","    at Worker.emit (node:domain:475:12)","    at Worker.[kOnExit] (node:internal/worker:281:10)","    at Worker.<computed>.onexit (node:internal/worker:201:20)"]}
/Users/Michael/workspace/theater-profiles/.sst/artifacts/mrooney-theatrum-profiles--core-stack-CreateProfileLambda/src/lambda/create-profile/handler.js:1871
          throw new Error("the worker has exited");
          ^

Error: the worker has exited
    at ThreadStream.flushSync (/Users/Michael/workspace/theater-profiles/.sst/artifacts/mrooney-theatrum-profiles--core-stack-CreateProfileLambda/src/lambda/create-profile/handler.js:1871:17)
    at process.onExit2 (/Users/Michael/workspace/theater-profiles/.sst/artifacts/mrooney-theatrum-profiles--core-stack-CreateProfileLambda/src/lambda/create-profile/handler.js:2084:16)
    at process.emit (node:events:406:35)
    at process.emit (node:domain:475:12)
    at process.exit (node:internal/process/per_thread:184:15)
    at process.<anonymous> (file:///Users/Michael/workspace/theater-profiles/node_modules/@serverless-stack/aws-lambda-ric/lib/index.js:31:17)
    at process.emit (node:events:394:28)
    at process.emit (node:domain:475:12)
    at process._fatalException (node:internal/process/execution:170:25)

https://serverless-stack.com/examples/how-to-use-middy-to-validate-your-serverless-api-requests.html Have you seen this example?

I have, yes. I’m using middy in a similar way but not sure what’s causing this strange “worker has exited” error.

I guess it’s related to internal nodejs threads.

@Michael Rooney just checking in to see if you had a chance to get middy to work?

No, I still haven’t been able to get it to work. I’m not sure what’s causing these internal node thread problems.

hmm.. which middy plugin are u using? Would you able to create a simple repo with this issue? And I can give it a try on my end.

I’m using multiple plugins. Here’s the helper function I wrap all the lambdas with:

import middy from "@middy/core";
import { Handler as LambdaHandler } from "aws-lambda";
import httpEventNormalizer from "@middy/http-event-normalizer";
import httpHeaderNormalizer from "@middy/http-header-normalizer";
import httpSecurityHeaders from "@middy/http-security-headers";
import jsonBodyParser from "@middy/http-json-body-parser";
import inputOutputLogger from "@middy/input-output-logger";
import httpErrorHandler from "@middy/http-error-handler";
import { log } from "../utils/logger";
import cors from "@middy/http-cors";
import validator from "@middy/validator";

export function with_middleware({ validators, logOptions }: TMiddlewareOptions = {}): (
  baseHandler: LambdaHandler
) => middy.MiddyfiedHandler {
  return (baseHandler) => {
    const lambda = middy(baseHandler);

    lambda
      .before(async (request) => {
        // Prevents jsonBodyParser from complaining on GET requests
        if (!request.event?.body) {
          request.event.body = "{}";
        }
      })
      .use(loggerMiddleware(logOptions)) // my own middleware that sets up a pino logger
      .use(inputOutputLogger({
        logger: (request) => {
            const child = log.child(request.context)
            <http://child.info|child.info>(request.event ?? request.response)
        },
        awsContext: true
      }))
      .use(httpSecurityHeaders()) // Applies best practice security headers to responses.
      .use(httpHeaderNormalizer({ canonical: true })) // Normalizes HTTP header names to their canonical format
      // Normalizes HTTP events by adding an empty object for queryStringParameters, multiValueQueryStringParameters or pathParameters if they are missing.
      .use(httpEventNormalizer({ payloadFormatVersion: 2 }))
      // Automatically parses HTTP requests with JSON body and converts the body into an object.
      // Also handles gracefully broken JSON if used in combination of httpErrorHandler.
      .use(jsonBodyParser())
      .use(httpErrorHandler())
      .use(cors())

    if (validators?.inputSchema) {
      lambda.use(
        validator({
          inputSchema: validators.inputSchema,
          ajvOptions: {
            strict: false,
            useDefaults: false,
          },
        })
      );
    }

    if (validators?.outputSchema) {
      lambda.use(
        validator({
          outputSchema: validators.outputSchema,
          ajvOptions: {
            strict: false,
            useDefaults: false,
          },
        })
      );
    }

    // lambda
    //   .use(identityProvider())
    //   .use(authenticate())
    //   .use(authorize())
    //   .use(cleanup());

    return lambda;
  };
}

I see. Let me give it a try.

Yes, those were always commented out since they're unused for now.

I’m getting a couple of type errors.

TMiddlewareOptions

and

loggerMiddleware

are not found.

Hmm. Interesting. I’m seeing a 10 second timeout error for “_flushSync took too long (10s)” now with this simple app I set up. I will put it into a git repo so you can pull it down and take a look.

Ah glad u figure it out!