NODE_OPTIONS='--insecure-http-parser' not being used with AWS Lambda - node.js

One of the web services my lambda connects to is periodically returning an invalid HTTP header and causing Node to throw with this error.
We are running lambda with the NodeJS 12.x runtime.
"name": "RequestError",
"message": "Error: Parse Error: Invalid header value char",
"cause": {
"bytesParsed": ,
"code": "HPE_INVALID_HEADER_TOKEN",
....
Following the instructions from this link it looks like the fix for 12.X would be to either use --insecure-http-parser or --http-parser=legacy as an option to starting Node.
Passing insecureHTTPParser to http.request() is not an option since we are using a http abstraction.
On Lambda the documented way to pass CLI options to use is to use the NODE_OPTIONS environment variable which is documented here
So I've tried setting either --insecure-http-parser or --http-parser=legacy and spawning a new lambda invocation and I'm still getting the HPE_INVALID_HEADER_TOKEN error.
Here's a screenshot of my Environment Variables in lambda.
Upgrading the NodeJS runtime is not an option too.
My high level question is how can I make this work?
Some other useful information in the comments would be
How can I verify that NODE_OPTIONS passed to Lambda in an environment variable are actually set at runtime?
How can I verify what HTTP parser is configured to be used in Node 12.x?

Related

[AWS][Amplify] Invoke function locally crashs with no error

I have just joined a developpment team, and the project should run in the cloud using amplify. I have a function called usershandler that i want to run locally. For that, i used :
amplify invoke function usershandler
This is the output i get :
Starting execution...
EVENT: {"httpMethod":"GET","body":"{\"name\": \"Amplify\"}","path":"/users","resource":"/{proxy+}","queryStringParameters":{}}
App started
get All VSM called
Connection to database was a success
null
Result:
{"statusCode":200,"body":"{\"success\":true,\"results\":[]}","headers":{"x-powered-by":"Express","access-control-allow-origin":"*","access-control-allow-headers":"Origin, X-Requested-With, Content-Type, Accept","content-type":"application/json; charset=utf-8","content-length":"29","etag":"W/\"1d-4wD7ChrrlHssGyekznKfKxR7ImE\"","date":"Tue, 21 Jul 2020 12:32:36 GMT","connection":"close"},"isBase64Encoded":false}
Finished execution.
EDIT : Also, when running the invoke command, amplify asks me for a src/event.json while i've seen it looking for the index.js for some ??
EDIT 2 [SOLVED] : downgrading #aws-amplify/cli to 4.14.1 seems to solve this :)
Expected behavior : The server should continue running so i can use it ..
Actual behavior : It always stops after the finished execution message.
The connection to the db works fine, the config.json contains correct values. Don't know why it is acting like this. Have anybody had the same problem?
Have a nice day.

Short answer: You are running the invoke command which is doing just what it is supposed to be doing - invoking the lambda function.
If you are looking to get a local API up, then run the following command:
sam local start-api
This will read your template and based on the endpoints you have setup, run them locally essentially mocking API Gateway locally. Read more about it in the official docs here.
Explanation:
This command comes is one of offering of AWS Serverless Application Model (AWS SAM). A tool to develop serverless application. It is essentially an abstraction of AWS Cloufdformation. Similarly Amplify is an abstraction that makes it simple to not only develop and manage the backend but also brings that power to frontend.
As both of them essentially use Cloudformation templates underneeth, you can leverage the capabilities of one tool with another.
SAM provides a robust set of tools for local development invcluding running a local lambda mocking server, in case you are not using API Gateway.
I use this combination to develop and test my frontend along with backend which is in golang, a language which is not as mature as javascript as a backend language with Amplify as of now.

How to set ssm param locally for serverless offline

I recently starting working on serverless architecture. Here is example of serverless.xml for the same.
test:
name: test
handler: handler.lambda_handler
timeout: 6
environment:
APP_ID: ${ssm:/path/to/ssm/test~true}
Now when I am trying to run serverless offline command then it complains about ssm variable.
Following is the error that coming on console.
I want to run everything on my locally machine for development. Can someone help on this how I can solve this problem.
ServerlessError: Trying to populate non string value into a string for variable ${ssm:/path/to/ssm/test~true}. Please make sure the value of the property is a string.
at Variables.populateVariable (C:\Users\kumarn\AppData\Roaming\npm\node_modules\serverless\lib\classes\Variables.js:464:13)
at Variables.renderMatches (C:\Users\kumarn\AppData\Roaming\npm\node_modules\serverless\lib\classes\Variables.js:386:21)
at C:\Users\kumarn\AppData\Roaming\npm\node_modules\serverless\lib\classes\Variables.js:406:29
From previous event:

you can solve this by adding the plugin:
https://github.com/janders223/serverless-offline-ssm
if you're feeling more adventurous you can also use localstack https://github.com/localstack/localstack
note that free version does not support everything

problem running simple demo example google-maps-services-js with Node

I have a problem running this simple demo example for google-maps-service with Node:
https://github.com/jpoehnelt/google-maps-services-js-demo
I am not sure what I am missing. When I run npm start, I get this error:
TypeError: Cannot read property 'elevation' of undefined
at D:\node_practice\google-maps-services-js-demo\index.js:14:35
at processTicksAndRejections (internal/process/task_queues.js:97:5)
I am using nvm and my node version is 13.9.0. I tried it on two different machines (one windows with above node version, one ubuntu) and both got the same error.

it seems the response coming back is empty. your r.data is empty, which means results[0] is undefined, so is the reason you are getting Cannot read property 'elevation' of undefined.
Below is the actual exception coming back from the APIs
"You must use an API key to authenticate each request to Google Maps Platform APIs. For additional information, please refer to http://g.co/dev/maps-no-account"
Get the instruction of getting an API key from here https://developers.google.com/maps/documentation/directions/get-api-key
and use the same with your code.

Google Cloud API Environment Credentials on Node.js on Phusion Passenger server

I have the translation client library set up on my Windows home Node.js server and use set GOOGLE_APPLICATION_CREDENTIALS=[PATH] to instantiate the credentials when I start the machine. This works fine.
On the publicly hosted server I don't think the request gets as far as asking for credentials because a different error is thrown than the one saying The request is missing a valid API key. This would appear I forgot to set the credentials or defined the path incorrectly.
I've tried many variations of export GOOGLE_APPLICATION_CREDENTIALS="[PATH]" but the error I keep getting is:
UnhandledPromiseRejectionWarning: Error: Unexpected error determining execution environment: Unsuccessful response status code. Request failed with status code 404
at GoogleAuth.<anonymous> (/usr/home/.../public_nodejs/node_modules/google-auth-library/build/src/auth/googleauth.js:163:23)
at Generator.throw (<anonymous>)
at rejected (/usr/home/.../public_nodejs/node_modules/google-auth-library/build/src/auth/googleauth.js:20:65)
at process._tickCallback (internal/process/next_tick.js:68:7)
I think this is to do with setting the environment variable in Phusion Passenger, but it's odd to me that the error I get is different.

AWS Request Expired Error

I'm trying to execute a script in which an AWS security group is created using NodeJS. When I try and run the script I get an error saying:
Error { RequestExpired: Request has expired.
Has anyone seen this? What does it mean? I thought the time between my local machine and the AWS Host might be off so I set a new timezone in my script to make up for the difference but still same error.

Resources