Timer in Java/ATG - Web Services call - java

I have generated a request for web services. I need to do do a check on my call. If the response is not returned within 5 seconds, another request will be shooted.
Pseudo Code :
webServiceClass response = xyz.getData();
If the response is not obtained in 5 seconds - send another request CheckData() to web services.This should be done for a maximum of 5 times.
I need to do this without using threads.

Try something like this (not tested but should give you the idea):
final MultiThreadedHttpConnectionManager httpConnections = new MultiThreadedHttpConnectionManager();
final HttpConnectionManagerParams connParams = manager.getParams();
final HttpClient httpClient = new HttpClient(manager);
final int connectionTimeout = 5000;
connParams.setConnectionTimeout(connectionTimeout);
try
{
// your web service call goes here
}
catch(ConnectTimeoutException cte)
{
if (isLoggingError())
{
logError(cte.getMessage());
}
}
catch(IOException ioe)
{
if (isLoggingError())
{
logError(ioe.getMessage());
}
}
finally
{
// make sure we always release the connection
method.releaseConnection();
}

Related

Looking for ways to detect AWS Lambda timeouts(few seconds before timeout) in Java and to test the same

My current Lambda function is calling a 3rd party web service Synchronously.This function occasionally times out (current timeout set to 25s and cannot be increased further)
My code is something like:
handleRequest(InputStream input, OutputStream output, Context context) throws IOException {
try{
response = calling 3rd party REST service
}catch(Exception e){
//handle exceptions
}
}
1)I want to custom handle the timeout (tracking the time and handling few milli seconds before actual timeout) within my Lambda function by sending a custom error message back to the client.
How can I effectively use the
context.getRemainingTimeInMillis()
method to track the time remaining while my synchronous call is running? Planning to call the context.getRemainingTimeInMillis() asynchronously.Is that the right approach?
2)What is a good way to test the timeout custom functionality ?
I solved my problem by increasing the Lambda timeout and invoking my process in a new thread and timing out the Thread after n seconds.
ExecutorService service = Executors.newSingleThreadExecutor();
try {
Runnable r = () ->{
try {
myFunction();
} catch (Exception e) {
e.printStackTrace();
}
};
f = service.submit(r);
f.get(n, TimeUnit.MILLISECONDS);// attempt the task for n milliseconds
}catch(TimeoutException toe){
//custom logic
}
Another option is to use the
readTimeOut
property of the RestClient(in my case Jersey) to set the timeout.But I see that this property is not working consistently within the Lambda code.Not sure if it's and issue with the Jersey client or the Lambda.
You can try with cancellation token to return custom exceptions with lambda before timeout.
try
{
var tokenSource = new CancellationTokenSource(TimeSpan.FromMinutes(1)); // set timeout value
var taskResult = ApiCall(); // call web service method
while (!taskResult.IsCompleted)
{
if (tokenSource.IsCancellationRequested)
{
throw new OperationCanceledException("time out for lambda"); // throw custom exceptions eg : OperationCanceledException
}
}
return taskResult.Result;
}
catch (OperationCanceledException ex)
{
// handle exception
}

Process for waiting for method response without blocking other HTTP requests

I have this code for a HTTP GET method:
#Override
public Cert get(Representation entity) {
// ...
Cert cert = ct.fetchCertificate(Arrays.asList(domains));
return cert;
}
It's a certificate generator that returns the serialized Cert object into the client.
And inside the fetchCertificate method is a Thread.sleep method that causes the entire servlet/web application to halt until the method returns, causing the whole web application not able to response to additional HTTP requests:
// Poll for the challenge to complete.
try {
int attempts = 20;
while (challenge.getStatus() != Status.VALID && attempts-- > 0) {
LOG.info("CHALLENGE ATTEMPTS: " + attempts);
// Did the authorization fail?
if (challenge.getStatus() == Status.INVALID) {
throw new AcmeException("Challenge failed... Giving up.");
}
// Wait for a few seconds
Thread.sleep(5000L);
// Then update the status
challenge.update();
}
} catch(Exception e) {
// ...
}
In RxJava terms what is the best way to achieve this same functionality but without blocking the application.
You need to put this Whole HTTP request calling and sleep inside a secondary thread. It seems it blocks the main thread.

Interrupt if API call to payment processor takes over 60 seconds

I'm integrating with a payments processor and am trying to deal with the scenario where:
user clicks pay and a request is made to our server
our server makes a request to the payment processor
there is a significant delay on the payment processor side
after a certain threshold e.g. 60 seconds, we alert the user that their payment was unsuccessful
after 70 seconds the payment processor returns a successful response.
So I need to start an API call to the payment processor from within the HTTP call from the UI, then if it takes more than 60 seconds, end the HTTP call and return an error to the user, then if the API call to the payment processor eventually succeeds (say after 70 seconds), send an email to the admin team.
I'm thinking of something like this:
import javax.ws.rs.client.*;
import java.util.Timer;
import java.util.TimerTask;
...
boolean overThreshold = false;
int timeout = 60; // seconds
TimerTask task = new TimerTask() {
#Override
public void run() {
overThreshold = true;
// return a message to user here saying their payment could not be processed
}
};
new Timer(true).schedule(task, timeout * 1000);
Client client = ClientBuilder.newClient();
WebTarget webTarget
= client.target({url of payment processor});
Invocation.Builder builder = webTarget.request()
.header(HttpHeaders.CONTENT_TYPE, APPLICATION_JSON);
final Response response = builder.post(Entity.json(new Gson().toJson(request)));
if (overThreshold) {
// send alert email here
}
There are a few problems, e.g. the run() method has void return value, error with overThreshold as being accessed from a inner class. Is there a more elegant way of doing this?
Using Future.get(timeout) from an ExecutorService should handle this quite cleanly.
For example:
ExecutorService executor = Executors.newCachedThreadPool();
// ... set up builder as before ...
Future<Response> responseFuture = executor.submit(
() -> builder.post(Entity.json(new Gson().toJson(request))));
try {
Response response = responseFuture.get(timeout, TimeUnit.SECONDS);
// return normal response here
} catch (TimeoutException ex) {
executor.submit( () -> {
Response lateResponse = responseFuture.get();
// send overThreshold alert email here
// Dummy return - prefer Callable to Runnable here for exception handling
return null;
} );
// return a message to user here saying their payment could not be processed
}
The choice of ExecutorService could be tuned to fit, or equally be a shared thread pool elsewhere in the application.

App Engine URL fetch callback

I am relatively new to App Engine. I do not understand how to asynchronously make an HTTP request with Java. I would have thought that this was a very simple matter using Thread and Runnable. But it seems that App Engine does not permit their use.
public Hashtable someApiMethod(..) {
SomeEntity entity = new SomeEntity(..);
ObjectifyService.ofy().save().entity(entity).now();
makeSomeHttpRequest(entity);
return launchResponse;
}
My question is this: how do I implement the method makeSomeHttpRequest(..) such that it returns without waiting for the URLFetchService.fetchAsync to return. I have tried the following without success:
protected void makeSomeHttpRequest(SomeEntity entity) {
URLFetchService fetcher = URLFetchServiceFactory.getURLFetchService();
try {
URL url = new URL("https://www.example.com");
Future future = fetcher.fetchAsync(url);
HTTPResponse response = (HTTPResponse) future.get();
byte[] content = response.getContent();
ByteArrayOutputStream bos = new ByteArrayOutputStream();
bos.write(content);
String responseString = new String(bos.toByteArray());
int responseCode = response.getResponseCode();
// Here I will do something with the responseCode and responseString
if (responseCode == 200) entity.someValue = responseString;
} catch (IOException e) {
// handle this
} catch (InterruptedException e) {
// handle this
} catch (ExecutionException e) {
// handle this
}
}
Really what I am trying to do is perform this HTTP request without forcing the method someApiMethod to wait for the response.
Few things:
First. Future doesn't work this way. Method .get waits for result of Feature execution, so basically you're stopping current thread until other thread finishes its execution. You're making it synchronous, it doesn't make any sense. Usually you call .get much later, when all other work in current thread is finished
Second. Threads in Appengine are limited to current request, you have to complete all async processing during current request. So updating an entity in such way doesn't make much sense, it's still bounded to current request. I mean in your case makeSomeHttpRequest(entity); should work faster than return launchResponse;
What you really need is to send this data to TaskQueue and make processing of SomeEntity entity from there (but don't send entity itself, just send ID and load by id from queue task). Basically it's going to be a new request handler (servlet/controller/etc) that should load entity by id, execute makeSomeHttpRequest (synchronously) and return http status 200.
See TaskQueue docs: https://cloud.google.com/appengine/docs/java/taskqueue/
You need Push Queue most likely: https://cloud.google.com/appengine/docs/java/taskqueue/overview-push

Registering with a URL for asynchronous notifications?

I am trying to understand more about async notifications. I have a URL in the form of:
http://www.sample.com/AsyncNotify?sessionId=xxxxxx
Now if I call this URL with the sessionId, it is equivalent to registering for Asynchronous notifications. I am using Apache HTTP Commons library to do Http Post and Get. If that's the case, then how can I receive events from the server side? Do I have to forget this approach and use sockets instead? Currently, this is my approach:
HttpClient httpClient = new HttpClient;
String url = "http://www.sample.com/AsyncNotify?sessionId=xxxxxx"
GetMethod get = new GetMethod(url);
try {
httpClient.executeMethod(get);
//read the response
} catch(Exception e) {
}
What I was thinking was to establish a socket level connection inside a while loop and call a handler whenever it receives some data, but is there a better way to achieve this?
EDIT:
I've used xSocket to get to the following stage but the connection closes after 30 seconds:
try {
String _GETRequest = "/sample/notify";
HttpClientConnection con = new HttpClientConnection("10.0.0.23", 5050);
con.setConnectionTimeoutMillis(100000);
GetRequest request = new GetRequest(_GETRequest);
request.setParameter("id", id);
IHttpResponseHandler responseHandler = new AsyncHandler();
con.send(request, responseHandler);
org.xlightweb.client.HttpClient httpClient = new org.xlightweb.client.HttpClient();
request.setParameter("id", id);
con.send(request, responseHandler);
// Don't let the program terminate. In other words,
// wait for a message from the server
while(con.isOpen()) {};
if(!con.isOpen()) {
}
} catch (ConnectException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
Tomcat came out with a technology called Tomcat Comet ( http://tomcat.apache.org/tomcat-6.0-doc/aio.html ). It has also been used for the new Servlet 3.0 spec. This technology will allow you to do persistent HTTP connections through which you can push notifications to any clients.
There is also a technolgy called WebSockets that is part of HTML 5
( http://dev.w3.org/html5/websockets/ ) Of course it only works in a limied set of browsers for now. Probably should wait on this.
Of course the current way to do it to be technolgy backwards compatible (even if it sucks) is to poll the server periodically and get results that way.
Of course if everybody (clients and servers) are on a local network then probably something like RMI or even EJBs or JMS Pub/Sub would be best.
Here is a Comet tutorial http://www.ibm.com/developerworks/web/library/wa-cometjava/index.html and another one http://www.javaworld.com/javaworld/jw-03-2008/jw-03-asynchhttp-test.html

Categories