Basic HTTP Authentication for CloudFront with Lambda@Edge

lambda-basic-auth.js

'use strict';
exports.handler = (event, context, callback) => {

    // Get request and request headers
    const request = event.Records[0].cf.request;
    const headers = request.headers;

    // Configure authentication
    const authUser = 'user';
    const authPass = 'pass';

    // Construct the Basic Auth string
    const authString = 'Basic ' + new Buffer(authUser + ':' + authPass).toString('base64');

    // Require Basic authentication
    if (typeof headers.authorization == 'undefined' || headers.authorization[0].value != authString) {
        const body = 'Unauthorized';
        const response = {
            status: '401',
            statusDescription: 'Unauthorized',
            body: body,
            headers: {
                'www-authenticate': [{key: 'WWW-Authenticate', value:'Basic'}]
            },
        };
        callback(null, response);
    }

    // Continue request processing if authentication passed
    callback(null, request);
};

See my article on Medium for details.

Thanks Leonid for sharing this.

Lines 8 to 13 if moved before the function (before line 2) it will be parsed just once and cached in memory for next executions.

This should be in the official AWS docs. Extremely useful; thanks for sharing!

Agreed ! There is little documentation on Lambda Edge. Thank you for sharing the code

Imakarov did you finish full doc on how to setup this? I am having hard time implementing this getting errors when assigning role to lambda. Which role should i use and which policies i have to add that role?

@trinath-k

I had the same problem. Your Lambda has to be in N. Virginia (althought that might have changed).

Find the role you assigned to the Lambda Function, and edit the policy "Trust Relationship" to something like this:

{ "Version": "2012-10-17", "Statement": [ { "Effect": "Allow", "Principal": { "Service": [ "edgelambda.amazonaws.com", "lambda.amazonaws.com" ] }, "Action": "sts:AssumeRole" } ] }

This worked for me. AWS Documentation does not explain that very well.

Good Luck !

Hi, I have configured everything. But i still can't access file in my s3 bucket. I am trying to access a file in this s3 bucket. it's only listing the object. I wanted it be able to download or read.

This XML file does not appear to have any style information associated with it. The document tree is shown below.

inventory-endpoints


1000
false

inventory-endpoints-04.06.2018.yaml
2018-04-13T20:40:38.000Z
"b18f7a425f4be500caf2fc7f3107e77c"
10308
STANDARD

Is there a way to set this up on specific directories in the s3 bucket? i.e. domain.com allowed, domain.com/dir triggers auth

@iamwalker In CloudFront, under "Origin Settings", you can set the Origin Path. For example, I'm running an S3 bucket with dev, stage and prod folders. I have one CloudFront distribution for each one.

You should return the callback or wrap the second call to it into a else case.

Hi!

Inspired from this I made my own version which is using a shared cookie secret to allow application to use Authorization header for JWT Bearer tokens. See webscale-oy/aws-cloudfront-basic-auth for Cloudformation templates and documentation.

Super helpful. Thank you very much!

With serverless configuration: https://tracklify.com/blog/simple-basic-auth-in-aws-cloudfront-with-serverless

Can we pick the credentials from secret manager? can anyone help with the code?

Can we pick the credentials from secret manager? can anyone help with the code?

https://aws.amazon.com/blogs/networking-and-content-delivery/securing-and-accessing-secrets-from-lambdaedge-using-aws-secrets-manager/

Do anyone get a problem that in phone (ios) it just popup password panel serveral times when loading the website?

Similar problem here.

It seems the Authorization header is no longer read by Lambda or some other part of the chain?

I replaced line 17 with:

const body = 'Unauthorized, headers are ' + JSON.stringify(headers);

Test case (with some sensitive data replaced by [...]):

$ curl --user xxx:yyy https://my_endpoint
[...]
> GET / HTTP/2
> Host: [...]
> Authorization: Basic eHh4Onl5eQ==
< HTTP/2 401
< content-length: 341
< server: CloudFront
< date: Mon, 28 Sep 2020 16:40:50 GMT
* Authentication problem. Ignoring this.
< www-authenticate: Basic
[...]

Unauthorized, headers is {"host":[{"key":"Host","value":"[...]"}],"x-forwarded-for":[{"key":"X-Forwarded-For","value":"[...]"}],"user-agent":[{"key":"User-Agent","value":"Amazon CloudFront"}],"via":[{"key":"Via","value":"[...]"}]}* Closing connection 0

The headers variable has no key authorization.

@rkz - make sure that your trigger CloudFront Event is Viewer Request and not Origin Request. The auth headers are stripped out of Origin.

@nachmore that was it :) Thank you

Line 23 is broken - the realm parameter is mandatory! This is documented in RFC7617 section 2 and I found out because Python's urllib's AuthHandler classes do not recognise the header without realm=.

This is now also possible, cheaper and faster with the newly launched CloudFront Functions (https://gist.github.com/jeroenvollenbrock/94edbbc62adc986d6d6a9a3076e66f5b)

@rkz - make sure that your trigger CloudFront Event is Viewer Request and not Origin Request. The auth headers are stripped out of Origin.

I added the 'Authorization' header to the cache key and it now comes through to the 'Origin Request'. My issue is that I need both this header as well as the origin domain for my lambda, and I can't determine the domain from the 'Viewer Request' . Wish this stuff was documented better.