ptrthomas / karate-devicefarm-demo

Karate and AWS DeviceFarm Demo

This is a project that demonstrates the following:

• Karate for Web-Browser automation using the W3C WebDriver Protocol
• AWS DeviceFarm (for desktop browser testing) as a remote WebDriver instance
• The use of a custom Target implementation that abstracts the code to start a WebDriver session out of your test. It does the following:
  • uses the AWS DeviceFarm SDK to provision a DeviceFarm URL / instance
  • shapes the webDriverSession and other driver config.
  • the code can be viewed here: DeviceFarmTarget.java

Prerequisites:

• Follow the instructions here to create a DeviceFarm project on AWS and set up your AWS access and secret keys in your environment.
• Change the urn in karate-config.js to match the DeviceFarm URN of the DeviceFarm project.

Now you can run SearchRunner.java as a JUnit test.

karate-devicefarm-demo's Issues

timeout capabilities(aws:maxDurationSecs & aws:idleTimeoutSecs) are not reflecting in Karate

When I added the timeout capabilities(aws:maxDurationSecs & aws:idleTimeoutSecs) in DeviceFarmTarget.java as shown below its not getting reflected. Please see the devicefarm webdriver.log below
--------------------------DeviceFarmTarget.java---------------------------------
@OverRide
public Map<String, Object> start(Logger logger) {
logger.info("making request to aws for browser: {}", browserName);
logger.info(" sercurty key arn : {}", arn);
DeviceFarmClient client = DeviceFarmClient.builder().region(Region.US_WEST_2).build();
CreateTestGridUrlRequest request = CreateTestGridUrlRequest.builder()
.expiresInSeconds(86400)
.projectArn(arn)
.build();
// logger.info(" request: {}", request);
CreateTestGridUrlResponse response = client.createTestGridUrl(request);
// logger.info(" response: {}", response);
String webDriverUrl = response.url();
// logger.info("aws url provisioned: {}", webDriverUrl);
Map<String, Object> map = new HashMap();
map.put("type", driverType);
map.put("start", false);
map.put("webDriverUrl", webDriverUrl);
// this is needed because it can take a minute or two for the "desktop" to be provisioned by aws
map.put("httpConfig", Collections.singletonMap("readTimeout", 120000));
// refer: https://docs.aws.amazon.com/devicefarm/latest/testgrid/techref-support.html
Map<String, Object> session = new HashMap();
map.put("webDriverSession", session);
Map<String, Object> capabilities = new HashMap();
capabilities.put("browserName", browserName);
capabilities.put("aws:maxDurationSecs", 2400);
capabilities.put("aws:idleTimeoutSecs", 900);
session.put("capabilities", capabilities);

    // for some reason, both are needed for aws device farm
    session.put("desiredCapabilities", capabilities);    
    return map;
}

----------------------------Webdriver.log-------------------------------
"capabilities": {
"alwaysMatch": {
"goog:chromeOptions": {
"args": [
"user-data-dir=C:\Users\testnode\chrome_user_data",
"--enable-logging --v=1"
]
}
},
"aws:idleTimeoutSecs": 240,
"aws:maxDurationSecs": 2400,
"browserName": "chrome"
}

The AWS suggested was trying setting the timeout capabilities(aws:maxDurationSecs & aws:idleTimeoutSecs) inside alwaysMatch .However, In the above case its outside the alwaysMatch.
IF this is the case, how do we add this capabilities in the DeviceFarmTarget.java so that it reflects inside the alwaysMatch block. Please let me know if you have any suggestion.

DeviceFarmTarget.docx