AWS Elastic Beanstalk node app degraded, no errors - amazon-web-services

Question

I'm suffering for a while to make elastic beanstalk work with my node app. No steps are going easily. After solving the running out of memory problem, now it's returning 502 bad gateway, and no errors are shown in the logs, or in the events.
Here are the few last lines of the log file:
2023/01/26 18:35:22.614796 [INFO] Running command /bin/sh -c systemctl show -p ConsistsOf eb-app.target | cut -d= -f2
2023/01/26 18:35:22.625486 [INFO] web.service
2023/01/26 18:35:22.625641 [INFO] Executing instruction: RunConfigDeployPostDeployHooks
2023/01/26 18:35:22.625655 [INFO] Executing platform hooks in .platform/confighooks/postdeploy/
2023/01/26 18:35:22.625669 [INFO] The dir .platform/confighooks/postdeploy/ does not exist
2023/01/26 18:35:22.625673 [INFO] Finished running scripts in /var/app/current/.platform/confighooks/postdeploy
2023/01/26 18:35:22.625685 [INFO] Executing cleanup logic
2023/01/26 18:35:22.625789 [INFO] CommandService Response: {"status":"SUCCESS","api_version":"1.0","results":[{"status":"SUCCESS","msg":"Engine execution has succeeded.","returncode":0,"events":[{"msg":"Instance deployment completed successfully.","timestamp":1674758122625,"severity":"INFO"}]}]}
2023/01/26 18:35:22.625945 [INFO] Platform Engine finished execution on command: config-deploy
2023/01/26 18:37:09.894419 [INFO] Starting...
2023/01/26 18:37:09.894459 [INFO] Starting EBPlatform-PlatformEngine
2023/01/26 18:37:09.894483 [INFO] reading event message file
2023/01/26 18:37:09.894643 [INFO] Engine received EB command cfn-hup-exec
2023/01/26 18:37:09.989819 [INFO] Running command /bin/sh -c /opt/aws/bin/cfn-get-metadata -s arn:aws:cloudformation:eu-west-1:538453521205:stack/awseb-e-sp3bw3fa7a-stack/768078f0-9d7c-11ed-8201-0aa34aaab773 -r AWSEBAutoScalingGroup --region eu-west-1
2023/01/26 18:37:10.305828 [INFO] Running command /bin/sh -c /opt/aws/bin/cfn-get-metadata -s arn:aws:cloudformation:eu-west-1:538453521205:stack/awseb-e-sp3bw3fa7a-stack/768078f0-9d7c-11ed-8201-0aa34aaab773 -r AWSEBBeanstalkMetadata --region eu-west-1
2023/01/26 18:37:10.620926 [INFO] checking whether command bundle-log is applicable to this instance...
2023/01/26 18:37:10.620942 [INFO] this command is applicable to the instance, thus instance should execute command
2023/01/26 18:37:10.620946 [INFO] Engine command: (bundle-log)
2023/01/26 18:37:10.621008 [INFO] Executing instruction: GetBundleLogs
2023/01/26 18:37:10.621014 [INFO] Bundle Logs...
As you can see there are no fails, the last block is getting the logs that you're reading. I cannot see it trying to run npm start.
Here's my package.json:
{
"name": "test",
"version": "0.0.1",
"author": "me",
"license": "private",
"description": "to be written",
"main": "index.js",
"private": true,
"scripts": {
"build": "webpack",
"dev": "nodemon NODE_ENV=development",
"start": "node index.js"
},
// dependencies, devDependencies
"engines": {
"node": ">=14 <=17.2.0"
}
}
I also gave the PORT environment variable under Configuration > Software > Environment properties with value 5000.
In theory it should automatically look for start script and execute it, but I cannot see any tries for that in the above logs.
When I connect to the ec2 instance of ebs, and route to my app, and run npm start I get the following:
> test#0.0.1 start
> node index.js
Executing command aws secretsmanager get-secret-value --secret-id mysecret --output json --region eu-west-1
node:internal/modules/cjs/loader:1249
return process.dlopen(module, path.toNamespacedPath(filename));
^
Error: /var/app/current/node_modules/bcrypt/lib/binding/napi-v3/bcrypt_lib.node: invalid ELF header
When I run aws secretsmanager get-secret-value --secret-id mysecret --output json --region eu-west-1 manually, it gives the expected output, no errors.
Because of the running out of memory while installing node_modules issue stated above, I solved it by installing the dependencies locally, and deploying the whole thing to ebs. It was on a mac.
Anyone any clue?