I have a controller proxy api endpoint where it receives different request payloads which are intended to different services. This controller validates payload and adds few headers based on certain rules. In this current context, i do not want to parse the received response from upstream services. proxy method should simply stream response to downstream clients so that it can scale well without going into any memory issues when dealing with large response payloads.
I have implemented method like this:
suspend fun proxyRequest(
url: String,
request: ServerHttpRequest,
customHeaders: HttpHeaders = HttpHeaders.EMPTY,
): ResponseEntity<String>? {
val modifiedReqHeaders = getHeadersWithoutOrigin(request, customHeaders)
val uri = URI.create(url)
val webClient = proxyClient.method(request.method!!)
.uri(uri)
.body(request.body)
modifiedReqHeaders.forEach {
val list = it.value.iterator().asSequence().toList()
val ar: Array<String> = list.toTypedArray()
#Suppress("SpreadOperator")
webClient.header(it.key, *ar)
}
return webClient.exchangeToMono { res ->
res.bodyToMono(String::class.java).map { b -> ResponseEntity.status(res.statusCode()).body(b) }
}.awaitFirstOrNull()
}
But this doesn't seems to be streaming. When i try to download large file, it is complaining failed to hold large data buffer.
Can someone help me in writing reactive streamed approach?
This is what i have done finally.
suspend fun proxyRequest(
url: String,
request: ServerHttpRequest,
response: ServerHttpResponse,
customHeaders: HttpHeaders = HttpHeaders.EMPTY,
): Void? {
val modifiedReqHeaders = getHeadersWithoutOrigin(request, customHeaders)
val uri = URI.create(url)
val webClient = proxyClient.method(request.method!!)
.uri(uri)
.body(request.body)
modifiedReqHeaders.forEach {
val list = it.value.iterator().asSequence().toList()
val ar: Array<String> = list.toTypedArray()
#Suppress("SpreadOperator")
webClient.header(it.key, *ar)
}
val respEntity = webClient
.retrieve()
.toEntityFlux<DataBuffer>()
.awaitSingle()
response.apply {
headers.putAll(respEntity.headers)
statusCode = respEntity.statusCode
}
return response.writeWith(respEntity.body ?: Flux.empty()).awaitFirstOrNull()
}
Let me know if this is truly sending data downstream and flushing?
Your first code snippet fails with memory issues because it is buffering in memory the whole response body as a String and forwards it after. If the response is quite large, you might fill the entire available memory.
The second approach also fails because instead of returning the entire Flux<DataBuffer> (so the entire response as buffers), you're only returning the first one. This fails because the response is incomplete.
Even if you manage to fix this particular issue, there are many other things to pay attention to:
it seems you're not returning the original response headers, effectively changing the response content type
you should not forward all the incoming response headers, as some of them are really up to the server (like transfer encoding)
what happens with security-related request/response headers?
how are you handling tracing and metrics?
You could take a look at the Spring Cloud Gateway project, which handles a lot of those subtleties and let you manipulate requests/responses.
Related
I'm interacting with a webservice which on POST request answers with 302 containing address to created resource in the location header. To access the created resource I've to make a GET request to the provided location.
I want reactor.netty.http.client.HttpClient to handle the redirect flow for me.
This is my configuration:
import reactor.netty.http.client.HttpClient;
...
...
var nettyHttpClient = HttpClient.create()
.compress(true)
.followRedirect(true);
With the above configuration, the client will use same HTTP method for the redirected request as it did for the first request.
Given my use-case, is there a way to provide my own redirect strategy to the client for 3xx responses?
You could handle raw response and handle http status 302 with custom logic
var response = nettyHttpClient.post()
.uri("/test")
.response()
.flatMap(res -> {
if (res.status().equals(HttpResponseStatus.FOUND)) {
return nettyHttpClient.get()
.uri(res.responseHeaders().get(HttpHeaders.LOCATION))
.response();
}
return Mono.just(res);
});
I have code that calls a vendor API to do a formdata upload of a file by axios from inside an AWS Lambda. The call returns a 400 error. If I run the code locally using the same node version v14 it works. I want to capture both raw requests and compare them for differences. How do I capture both raw requests? I've tried using ngrok and pipedream but they don't show the raw but decode the request and the file.
let response = null;
try {
const newFile = fs.createReadStream(doc);
const formData = new FormData();
formData.append("file", newFile);
formData.append("url", url);
const headers = {
Authorization: "Bearer " + token,
...formData.getHeaders(),
};
console.log("Headers: ", headers);
response = await axios.post(`${APIBASE}/file/FileUpload`, formData, {
headers,
});
console.log("file upload response", response);
} catch (err) {
console.log("fileupload error at API", err);
}
You might be able to just use a custom request interceptor and interrogate at the requests that way.
https://axios-http.com/docs/interceptors
You're not able to capture the request on the network level, as this is totally controlled by AWS. Maybe there's a way to do this when running in a VPC, but I don't think so.
You could simply use a tool such as axios debug logger to print out all of the request and response contents (including headers etc) before the request is made/after the response has arrived. This might provide some more information as to where things are going wrong.
As to the cause of the problem, it is difficult to help you there since you haven't shared the error message nor do we know anything about the API you're trying to call.
There are multiple ways to debug
axios debug logger .
AWS cloud watch where you can see all the logs. you can capture the request
and response.
Use postman to call the prod lambda endpoint and verify the response.
In my application, it is important for the user to know when the data was updated for the last time.
For example, let's say that the response from a Web API is cached like this:
workbox.routing.registerRoute(
/https:\/\/api\.service\.com\/resource.*/,
workbox.strategies.cacheFirst({
cacheName: 'resource',
plugins: [
new workbox.expiration.Plugin({
maxAgeSeconds: ONE_HOUR,
}),
]
}),
)
During one hour, the response will come from cache, and it can be much longer if the user is offline.
How can I know the original response date/time?
You'd probably want to retrieve that information from the Date: header of the response. Unfortunately, if you're dealing with a CORS response from a third-party server, the Date: header is not available by default—only the simple response headers are available.
So the first thing you'd want to do is make sure that the API you're using sets Access-Control-Expose-Headers: Date on its HTTP responses.
I don't know how you're reading the API response in your web app, but let's assume that you're using fetch() to obtain the response (which may or may not come from the service worker's cache) and then processing the response as JSON. Code to examine the Date: header could look something like:
const response = await fetch('https://api.service.com/resource');
const dateAsString = response.headers.get('Date');
const dateAsMillis = Date.parse(dateAsString);
// Do something with dateAsMillis, like comparing it to Date.now().
if ((Date.now() - dateAsMillis) < SOME_THRESHOLD) {
const json = await response.json();
// Etc.
}
I'm using Spring Boot to create an API that needs to be consumed in Angular 4. Spring and Angular are on different ports.
The problem is that Spring's ResponseEntity raises an error in Angular.
#RequestMapping(value = "/{id}", method = RequestMethod.GET)
public ResponseEntity getFlow(#PathVariable int id) {
Flow flow = flowService.findById(id);
return new ResponseEntity(flow, HttpStatus.FOUND);
}
Now, I can perfectly use Postman to test the API and it works.
But when I make a request from Angular, it returns an error:
Strangely, it returns an error alongside the requested object.
Now, the cause of the problem is that the Spring Boot application returns a ResponseEntity and not a normal object (like String), and Angular doesn't know how to interpret it. If the controller returns just a Flow object, it works.
How can it be solved using ResponseEntity? Or, how else can I send the object alongside the HTTP status code?
Also, in #RequestMapping put produces = "application/json", and in get request in angular, add http options :
const httpOptions = {
headers: new HttpHeaders({
'Accept': 'application/json',
'Content-Type': 'application/json'
})
};
So your get request looks like this:
this.http.get(url, httpOptions)
As per the document mentioned here
https://docs.angularjs.org/api/ng/service/$http
A response status code between 200 and 299 is considered a success status and will result in the success callback being called. Any response status code outside of that range is considered an error status and will result in the error callback being called. Also, status codes less than -1 are normalized to zero. -1 usually means the request was aborted, e.g. using a config.timeout. Note that if the response is a redirect, XMLHttpRequest will transparently follow it, meaning that the outcome (success or error) will be determined by the final response status code.
As you are sending an instance of ResponseEntity(HttpStatus.Found) whose Http status code is 302 which doesnt fall under the success range thats why error callback is called.
Try returning the content like this
return new ResponseEntity(flow, HttpStatus.OK);
Currently my client authenticates request only on case of 401 response:
this.client.authenticator(new okhttp3.Authenticator() {
public Request authenticate(Route route, Response response) throws IOException {
String credentials = authenticator.getCredentials();
if (credentials.equals(response.request().header("Authorization"))) {
throw new TraversonException(401, "Unauthorized", response.request().url().toString());
} else {
defaultHeader("Authorization", credentials);
Request.Builder newRequest = response.request().newBuilder()
.headers(Headers.of(defaultHeaders));
return newRequest.build();
}
});
But I'd like to change this behavior and be able to call it either manually or auto per first call? Is it possible somehow?
If the authentication is predictably required and not related to a proxy, then the solution is to implement an Interceptor instead of Authenticator.
OkHttpClient.Builder clientBuilder = ...;
clientBuilder.networkInterceptors().add(0, myInterceptor);
client = clientBuilder.build();
Example Interceptor https://github.com/yschimke/oksocial/blob/48e0ca53b85e608443eab614829cb0361c79aa47/src/main/java/com/baulsupp/oksocial/uber/UberAuthInterceptor.java
n.b. There is discussion around possible support for this usecase in https://github.com/square/okhttp/pull/2458. One issue with current Authenticator API is that it assumes a Response from the failed (401) request.