Disable retry / keep-alive in Rest Assured for long responsetimes (no retry every 60 sec)

11/26/2019

I have an issue that Rest-Assured seems to issue a 2nd and 3rd call when the response takes more than 60 seconds. What I need is RestAssured to wait for the call to finish and not send more calls.

The simple test to an endpoint that takes more than 60 seconds to respond:

        return given()
                .relaxedHTTPSValidation()
                .baseUri(cloudBaseurl)
                .when()
                .get(fullServicePath)
                .then().extract().response();

The issue is that I see in the logging of the requested service that exactly 60 seconds later, a second call comes in and a 3rd after another 60 seconds. By default RestAssured seems to timeout after 3 minutes.

What I tried:

1. Setting different parameters

        RestAssured.config= RestAssuredConfig.config().httpClient(httpClientConfig().
                setParam("http.connection.timeout",70000).
                setParam("http.connection.request.timeout",70000).
                setParam("http.socket.timeout",70000).
                setParam("http.connection-manager.timeout",70000).
                setParam("http.conn-manager.timeout",70000L).
                setParam("http.connection.stalecheck",false).
                setParam("http.keepAlive",70000L));

It does timeout after 70 seconds, but still after 60 seconds the 2nd call comes in.

2. Setting headers in RestAssured

        return given()
                .relaxedHTTPSValidation()
                .baseUri(cloudBaseurl)
                .header(HttpHeaders.CONNECTION,"Keep-Alive")
                .header("Keep-Alive","timeout=100", "max=180")
                .when()
                .get(fullServicePath)
                .then().extract().response();

No change. Still a 2nd request after 60 seconds.

3. Changing the default request behavior (as suggested here: https://stackoverflow.com/questions/23054289/httpclient-executes-requests-multiple-time-if-request-timed-out)

        RestAssured.config = RestAssured.config().httpClient(httpClientConfig().httpClientFactory(
                () -> {
                    SystemDefaultHttpClient systemDefaultHttpClient = new SystemDefaultHttpClient();
                    ClientConnectionManager connectionManager = systemDefaultHttpClient.getConnectionManager();
                    connectionManager.closeIdleConnections(100, TimeUnit.SECONDS);
// Disable default behavior of HttpClient of retrying requests in case of failure
                    ((AbstractHttpClient) systemDefaultHttpClient).setHttpRequestRetryHandler(new DefaultHttpRequestRetryHandler(0, false));
                    return systemDefaultHttpClient;
                }));
-- Blitzoff
apache-httpclient-4.x
kubernetes
kubernetes-ingress
rest-assured

1 Answer

11/26/2019

I was looking in the wrong direction.

The timeout on the Ingress in Kubernetes needed to be increased. Adding the following to the Kubernetes deployment yaml solved the issue.

metadata:
  name: your-service-name
  labels:
    app: your-service-name
  annotations:
    kubernetes.io/ingress.class: "nginx"
    nginx.ingress.kubernetes.io/proxy-connect-timeout : "3600"
    nginx.ingress.kubernetes.io/proxy-read-timeout : "3600"
    nginx.ingress.kubernetes.io/proxy-send-timeout : "3600"
-- Blitzoff
Source: StackOverflow