Multiple Kubernetes authorization modules checked in sequence, how? - kubernetes

From the Kubernetes documentation on authorization it states that:
When multiple authorization modules are configured, each is checked in sequence. If any authorizer approves or denies a request, that decision is immediately returned and no other authorizer is consulted. If all modules have no opinion on the request, then the request is denied. A deny returns an HTTP status code 403.
I am now writing a custom webhook for authorization and I would want the logic to fallback to RBAC for a few cases - i.e. have my webhook respond with what the documentation refers to as "no opinion". The documentation however only details how to approve or deny a request and doesn't come back to this third option which seems essential for having multiple authorization modules checked in sequence. How would I best in the context of my webhook respond with "I have no opinion on this request, please pass it on to the next authorizer"?

It's not clear how multiple AuthorizationModule work from kubernetes official doc.
So I check the source code of apiserver, it create a combine authorizer.Authorizer by union.New(authorizers...), from union source I find the answer:
The union authorizer iterates over each subauthorizer and returns the first decision that is either an Allow decision or a Deny decision. If a subauthorizer returns a NoOpinion, then the union authorizer moves onto the next authorizer or, if the subauthorizer was the last authorizer, returns NoOpinion as the aggregate decision
More detail at k8s.io/apiserver/pkg/authorization/union:
func (authzHandler unionAuthzHandler) Authorize(a authorizer.Attributes) (authorizer.Decision, string, error) {
var (
errlist []error
reasonlist []string
)
for _, currAuthzHandler := range authzHandler {
decision, reason, err := currAuthzHandler.Authorize(a)
if err != nil {
errlist = append(errlist, err)
}
if len(reason) != 0 {
reasonlist = append(reasonlist, reason)
}
switch decision {
case authorizer.DecisionAllow, authorizer.DecisionDeny:
return decision, reason, err
case authorizer.DecisionNoOpinion:
// continue to the next authorizer
}
}
return authorizer.DecisionNoOpinion, strings.Join(reasonlist, "\n"), utilerrors.NewAggregate(errlist)
}
So if you want to create your custom webhook AuthozitaionModule, if you want to pass decision to next authorizer, just give permissive response like:
{
"apiVersion": "authorization.k8s.io/v1beta1",
"kind": "SubjectAccessReview",
"status": {
"reason": "no decision",
"allowed": false,
"denied": false
}
}
Then apiserver can make a decision by this reponse:
switch {
case r.Status.Denied && r.Status.Allowed:
return authorizer.DecisionDeny, r.Status.Reason, fmt.Errorf("webhook subject access review returned both allow and deny response")
case r.Status.Denied:
return authorizer.DecisionDeny, r.Status.Reason, nil
case r.Status.Allowed:
return authorizer.DecisionAllow, r.Status.Reason, nil
default:
return authorizer.DecisionNoOpinion, r.Status.Reason, nil
}

Related

How to make PUT request RESTful?

With a single API resource /, we have written only one handler that process GET & POST request on API resource /
POST we use to create a resource in database, byt sending data in request body
PUT we use to update an existing resource in database
My understanding is, RESTful best practice says, a handler need to serve an API resource(say /) for all requests GET, POST & PUT
We want the same handler to process PUT request, but the API resource will be something like /1234, where 1234 is existing id
Technically, API resource /1234 will also map to same handler that processes /, but,
From RESTful best practices, Does /1234 need to be handled without passing id as part of API resource URI? something like below...
func ServeHTTP(w http.ResponseWriter, r *http.Request) {
if r.Method == http.MethodGet { // for API resource '/'
p.getProducts(w, r)
return
}
if r.Method == http.MethodPost { // for API resource '/'
p.addProduct(w, r)
return
}
if r.Method == http.MethodPut { // for API resource '/'
p.updateProduct(w, r)
return
}
}
func updateProduct(w http.ResponseWriter, r *http.Request) {
var idString string
decoder := json.NewDecoder(r.Body)
decoder.Decode(idString)
id, err := findID(idString)
// do whatever with id
}
func findID(str string) (int, error) {
dfa := regexp.MustCompile(`/([0-9]+)`)
matches := dfa.FindAllStringSubmatch(str, -1) // returns [][]string
idString := matches[0][1]
id, err := strconv.Atoi(idString)
return id, nil
}
As I understood right you right.
You have two call which can be handle without Id for end point /.
One is POST when the back-end with generate you Id as a result.
Second is GET for all resources but this is up to you. Maybe because of secure reason you would not like to list all available resources.
One extra information is that PUT & 'POST' can use the same handler but logic in handler has to check if 'id' is provided and do extra more logic to create resource.

Firestore PHP SDK using Auth Rule setup

I am new to Firebase, right now i want to try Firestore php SDK and implement firestore auth rule. Current code below is work fine
use Google\Cloud\Firestore\FirestoreClient;
$db = new FirestoreClient();
$db->collection('mycollectionname')
->document('mydocumentname')
->set(['name'=>'aaa','value'=>'111');
Firestore auth rule
rules_version = '2';
service cloud.firestore {
match /databases/{database}/documents {
match /{document=**} {
// before change
allow read, write: if true;
// after change
allow read, write: if request.auth.uid != null;
}
}
}
After change true to request.auth.uid != null, its give me error:
{ "error": { "code": 403, "message": "Missing or insufficient permissions.", "status": "PERMISSION_DENIED" } }
I can figure out to get user data like : email & password or user id token, how to solve above error using user data?
I just found it, maybe useful for anyone.
on file use Google\Cloud\Firestore\Connection\Grpc
I made some change :
......
......
private function addRequestHeaders(array $args)
{
$args += [
'headers' => []
];
$args['headers']['google-cloud-resource-prefix'] = [$this->resourcePrefixHeader];
///////// CODE THAT I ADD ///////////
if(session('user_token'))
{
$args['headers']['Authorization'] = ['Bearer '.session('user_token')];
}
///////// CODE THAT I ADD ///////////
// Provide authentication header for requests when emulator is enabled.
if ($this->isUsingEmulator) {
$args['headers']['Authorization'] = ['Bearer owner'];
}
return $args;
}
Using laravel, I made it to check the session, to add a Bearer token on header. After that using example above i added a session before the Firestore function was used
use Google\Cloud\Firestore\FirestoreClient;
/// adding session
session(['user_token'=>'eyJhbGciOiJSUzI1................']);
$db = new FirestoreClient();
$db->collection('users')
->document('test#gmail.com')
->set(['name'=>'aaa','value'=>'111');
In the auth rule I can also make document rules according to the user's email name, where the user's email can be obtained with the token id that was added before. Example:
rules_version = '2';
service cloud.firestore {
match /databases/{database}/documents {
match /users/{email} {
allow read, write: if email == request.auth.token.email;
}
}
}
Hellow,
If you want implement FirestoreClient PHP Library in your backend, you should use a service account to allow your backend go through firestore rules, this way you can control users flow (with rules) and give complete access or limited access to your own backend code. You can see how to set up a service account in Google's quicstart documentation.
In this process you might have some issue while trying to configure the path to your service account key file. Which might throw this error:
Google\Cloud\Core\Exception\ServiceException:
{
"message": "Missing or insufficient permissions.",
"code": 7,
"status": "PERMISSION_DENIED"
}
If this happens to you see my response in this stackoverflow thread.
I hope it helps you and everyone having the same issue.

How to decorate Siesta request with an asynchronous task

What is the correct way to alter a Request performing an asynchronous task before the Request happens?
So any request Rn need to become transparently Tn then Rn.
A little of background here: The Task is a 3rd party SDK that dispatch a Token I need to use as Header for the original request.
My idea is to decorate the Rn, but in doing this I need to convert my Tn task into a Siesta Request I can chain then.
So I wrapped the Asynchronous Task and chained to my original request.
Thus any Rn will turn into Tn.chained { .passTo(Rn) }
In that way, this new behaviour is entirely transparent for the whole application.
The problem
Doing this my code end up crashing in a Siesta internal precondition:
precondition(completedValue == nil, "notifyOfCompletion() already called")
In my custom AsyncTaskRequest I collect the callbacks for success, failure, progress etc, in order to trigger them on the main queue when the SDK deliver the Token.
I noticed that removing all the stored callback once they are executed, the crash disappear, but honestly I didn't found the reason why.
I hope there are enough informations for some hints or suggests.
Thank you in advance.
Yes, implementing Siesta’s Request interface is no picnic. Others have had exactly the same problem — and luckily Siesta version 1.4 includes a solution.
Documentation for the new feature is still thin. To use the new API, you’ll implement the new RequestDelegate protocol, and pass your implementation to Resource.prepareRequest(using:). That will return a request that you can use in a standard Siesta request chain. The result will look something like this (WARNING – untested code):
struct MyTokenHandlerThingy: RequestDelegate {
// 3rd party SDK glue goes here
}
...
service.configure(…) {
if let authToken = self.authToken {
$0.headers["X-Auth-Token"] = authToken // authToken is an instance var or something
}
$0.decorateRequests {
self.refreshTokenOnAuthFailure(request: $1)
}
}
func refreshTokenOnAuthFailure(request: Request) -> Request {
return request.chained {
guard case .failure(let error) = $0.response, // Did request fail…
error.httpStatusCode == 401 else { // …because of expired token?
return .useThisResponse // If not, use the response we got.
}
return .passTo(
self.refreshAuthToken().chained { // If so, first request a new token, then:
if case .failure = $0.response { // If token request failed…
return .useThisResponse // …report that error.
} else {
return .passTo(request.repeated()) // We have a new token! Repeat the original request.
}
}
)
}
}
func refreshAuthToken() -> Request {
return Request.prepareRequest(using: MyTokenHandlerThingy())
.onSuccess {
self.authToken = $0.jsonDict["token"] as? String // Store the new token, then…
self.invalidateConfiguration() // …make future requests use it
}
}
}
To understand how to implement RequestDelegate, you best bet for now is to look at the new API docs directly in the code.
Since this is a brand new feature not yet released, I’d greatly appreciate a report on how it works for you and any troubles you encounter.

How can I change the bearer token in Moya

the documentation shows how to make targets require bearer tokens, which I did like this
extension MyService: AccessTokenAuthorizable {
var authorizationType: AuthorizationType {
switch self {
case .resetPassword, .postTextBook, .bookmarkBook, .getBookmarks, .logout, .verify:
return .bearer
default:
return .none
}
}
}
then it shows how to add tokens to the providers, which I did like this
let token = "abc123"
let authPlugin = AccessTokenPlugin(tokenClosure: token)
let provider = MoyaProvider<MyService>(plugins: [authPlugin])
but when the token expires, how can I change the token? and does Moya offer a way to automate this process, where if I get a forbidden http response (meaning I am not authorized), it automatically requests a token?
The implementation details of authentication/authorization can be quite different for each API out there. This is the reason why Moya will not handle the auth for you.
That said, implementing your own authentication/authorization can be done in many ways. It will depend on your constraints and/or preferences. As of today, you can find a few solutions sparsely outlined in Moya documentation:
Use the PluginType to add your auth to the requests. But think that this can potentially be used to refresh the token if needed. You may also need to intercept the completion of the request to detect authorization errors and apply your preferred recovery scenario (eg. refresh the token and retry the call).
Same can be implemented using the endpointClosure and/or requestClosure.
You can also consider implementing Alamofire's RequestAdapter and RequestRetrier. Depending on your needs, this can make retries easier. However, on them you will not have straightforward access to your TargetType, so you may need to find a way to recognize the different auth methods needed (ie. your bearer or none).
A few direct references to their documentation:
Plugins
Endpoints
Authentication
Alamofire Automatic Validation
Also, I highly encourage anybody to learn/get inspiration from Eilodon's Networking source code.
for change/refresh token i used this
static func send(request: TargetType) -> PrimitiveSequence<SingleTrait, Response> {
return provider.rx.request(request)
.retry(1)
.observeOn(ConcurrentDispatchQueueScheduler.init(qos: .default))
.filterSuccessfulStatusAndRedirectCodes()
.retryWhen({ (errorObservable: Observable<Error>) in
errorObservable.flatMap({ (error) -> Single<String> in
if let moyaError: MoyaError = error as? MoyaError, let response: Response = moyaError.response {
if **check forbidden http responses here** {
return provider.rx.request(.refreshToken(*your refresh token here*))
.filterSuccessfulStatusCodes()
.mapString(atKeyPath: "*json path to new access token*")
.catchError { (_) in
logout()
throw error
}
.flatMap({ (newAccessToken) -> PrimitiveSequence<SingleTrait, String> in
changeAccessToken()
return Single.just(newAccessToken)
})
}
}
throw error
})
})
}
static func logout() {
// logout action
}
static func changeAccessToken() {
// set new access token
}

How to throw custom error message from API Gateway custom authorizer

Here in the blue print says, API gateway will respond with 401: Unauthorized.
I wrote the same raise Exception('Unauthorized') in my lambda and was able to test it from Lambda Console. But in POSTMAN, I'm receiving status 500
with body:
{
message: null`
}
I want to add custom error messages such as "Invalid signature", "TokenExpired", etc., Any documentation or guidance would be appreciated.
This is totally possible but the docs are so bad and confusing.
Here's how you do it:
There is an object called $context.authorizer that you have access to in your gateway responses template. You can read more about it here: https://docs.aws.amazon.com/apigateway/latest/developerguide/api-gateway-mapping-template-reference.html
Here is an examample of populating this authorizer object from your authorizer lambda like so:
// A simple TOKEN authorizer example to demonstrate how to use an authorization token
// to allow or deny a request. In this example, the caller named 'user' is allowed to invoke
// a request if the client-supplied token value is 'allow'. The caller is not allowed to invoke
// the request if the token value is 'deny'. If the token value is 'Unauthorized', the function
// returns the 'Unauthorized' error with an HTTP status code of 401. For any other token value,
// the authorizer returns an 'Invalid token' error.
exports.handler = function(event, context, callback) {
var token = event.authorizationToken;
switch (token.toLowerCase()) {
case 'allow':
callback(null, generatePolicy('user', 'Allow', event.methodArn));
break;
case 'deny':
callback(null, generatePolicy('user', 'Deny', event.methodArn));
break;
case 'unauthorized':
callback("Unauthorized"); // Return a 401 Unauthorized response
break;
default:
callback("Error: Invalid token");
}
};
var generatePolicy = function(principalId, effect, resource) {
var authResponse = {};
authResponse.principalId = principalId;
if (effect && resource) {
var policyDocument = {};
policyDocument.Version = '2012-10-17';
policyDocument.Statement = [];
var statementOne = {};
statementOne.Action = 'execute-api:Invoke';
statementOne.Effect = effect;
statementOne.Resource = resource;
policyDocument.Statement[0] = statementOne;
authResponse.policyDocument = policyDocument;
}
// Optional output with custom properties of the String, Number or Boolean type.
authResponse.context = {
"stringKey": "stringval custom anything can go here",
"numberKey": 123,
"booleanKey": true,
};
return authResponse;
}
They key here is adding this part:
// Optional output with custom properties of the String, Number or Boolean type.
authResponse.context = {
"stringKey": "stringval custom anything can go here",
"numberKey": 123,
"booleanKey": true,
};
This will become available on $context.authorizer
I then set the body mapping template in gateway responses tab like this:
{"message":"$context.authorizer.stringKey"}
NOTE: it must be quoted!
finally - after sending a request in postman with Authorization token set to deny I now get back a payload from postman that looks like this:
{
"message": "stringval custom anything can go here"
}
I used #maxwell solution, using custom resource ResponseTemplates. For deny response see below:
{
"success":false,
"message":"Custom Deny Message"
}
You can check this out here: https://github.com/SeptiyanAndika/serverless-custom-authorizer
I'm not sure what is causing the 500 message: null response. Possibly misconfiguration of the Lambda function permissions.
To customize the Unauthorized error response, you'll set up a Gateway Response for the UNAUTHORIZED error type. You can configure response headers and payload here.
Maxwell is mostly correct. I tried his implementation and noticed that his message should go from :
{"message":"$context.authorizer.stringKey"}
to
{"message":"$context.authorizer.context.stringKey"}
As noted by Connor far as I can see, the answer to the specific question - which mentions 401 related errors - is NO.
You can produce a generic 401 Unauthorized but you cannot alter the error message.
That is you can customise the 403 Forbidden (DENY) messages but not the 401's.
Note that I've used the NodeJS Lambda custom authorizers but not the Python version referenced in the question.
With my testing what i observed is , You cannot customize message when you throw exception from the lambda,
You can have customized messages when you return DENY Policy message from the authorizer
Here is how i am returning custom message when i DENY from the Authorizer, it in the detail field of
authResponse.context returned from custom Authorizer
you can also update status code to 401 instead of 403 .
This can be easily achieved by using the context.fail() function.
Example:
const customAuthorizer: Handler = (event, context: Context, callback: Callback) => {
authenticate(event)
.then((res) => {
// result should be as described in AWS docs
callback(null, res);
})
.catch((err) => {
context.fail("Unauthorized");
});
}
This will return a 401 response with following body.
{
"message": "Unauthorized"
}
This can also be achieved by throwing an error:
throw new Error('Unauthorized');