Run Selenium Chrome WebDriver on Azure Cloud Service?

Question

I have : ASP.NET Core2 App + Selenium to automate some actions with browser.

It works perfect on local.Use the latest versions of all nuget and exe.

After deploy to Azure have problems on create Webdriver.

I tried:

• Include .exe files to folder and use it like :

new ChromeDriver(ChromeDriverService.CreateDefaultService("./CORE/ExeFiles"), chromeOptions);

• Run Job with standalone RemoteWebServer: Cant connect to it + Job disappears after Start-Stop site.
• Run .exe files as service - dead end;
• Run .exe file from CMD with code: RemoteWebServer on 4444 port OK, but I can't connect to it.

Read about some Firewall or Antivirus blocking stuff but cant find where to configure necessary properties on Azure.

How can I use Selenium on Azure? Some simplest example pls??, I'm fighting with this for 3 days =(

P.S. Also find this article https://github.com/projectkudu/kudu/wiki/Azure-Web-App-sandbox#unsupported-frameworks and THIS in the end:

Unsuported: PhantomJS/Selenium: tries to connect to local address, and also uses GDI+.

Alternatives? How to use Selenium on Azure?

Answer 1

string apikey = ConfigurationManager.AppSettings["BROWSERLESS_API_KEY"];
ChromeOptions chromeOptions = new ChromeOptions();
// Set launch args similar to puppeteer's for best performance
chromeOptions.AddArgument("--disable-background-timer-throttling");
chromeOptions.AddArgument("--disable-backgrounding-occluded-windows");
chromeOptions.AddArgument("--disable-breakpad");
chromeOptions.AddArgument("--disable-component-extensions-with-background-pages");
chromeOptions.AddArgument("--disable-dev-shm-usage");
chromeOptions.AddArgument("--disable-extensions");
chromeOptions.AddArgument("--disable-features=TranslateUI,BlinkGenPropertyTrees");
chromeOptions.AddArgument("--disable-ipc-flooding-protection");
chromeOptions.AddArgument("--disable-renderer-backgrounding");
chromeOptions.AddArgument("--enable-features=NetworkService,NetworkServiceInProcess");
chromeOptions.AddArgument("--force-color-profile=srgb");
chromeOptions.AddArgument("--hide-scrollbars");
chromeOptions.AddArgument("--metrics-recording-only");
chromeOptions.AddArgument("--mute-audio");
chromeOptions.AddArgument("--headless");
chromeOptions.AddArgument("--no-sandbox");
chromeOptions.AddAdditionalCapability("browserless.token", apikey, true);
using (var driver = new RemoteWebDriver(new Uri("https://chrome.browserless.io/webdriver"), chromeOptions.ToCapabilities()))
{
//Your selenium code
}

Answer 2

It won't work on App Service and you already found the limits and limitations page that explains it.

That being said, it works just fine on a Cloud Service (with Roles), yes the good ol' Cloud Services.

WebRole sample —

using System;
using System.Collections.Generic;
using System.Linq;
using System.Net;
using System.Web.Http;
using OpenQA.Selenium;
using OpenQA.Selenium.PhantomJS;
using OpenQA.Selenium.Support.UI;

namespace WebRole1.Controllers
{
    public class PhantomController : ApiController
    {
        /// <summary>
        /// Run PhantomJS UI tests against the specified URL
        /// </summary>
        /// <param name="URL">URL to test</param>
        public string Get(string URL)
        {
            string result = UITests.Test(URL);
            return result;
        }
    }

    /// <summary>
    /// UITests class
    /// </summary>
    public class UITests
    {
        /// <summary>
        /// Test implementation
        /// </summary>
        /// <param name="URL">URL to test</param>
        /// <returns></returns>
        public static string Test(string URL)
        {
            // Initialize the Chrome Driver
            // Place phantomjs.exe driver in the project root,
            // meaning same folder as WebRole.cs
            using (var driver = new PhantomJSDriver())
            {
                try
                {
                    // Go to the home page
                    driver.Navigate().GoToUrl(URL);

                    IWebElement input;
                    WebDriverWait wait = new WebDriverWait(
                        driver, TimeSpan.FromSeconds(2));

                    Func<IWebDriver, IWebElement> _emailInputIsVisible =
                        ExpectedConditions.ElementIsVisible(By.Id("email"));
                    wait.Until(_emailInputIsVisible);
                    input = driver.FindElementById("email");
                    input.SendKeys("[email protected]");
                    driver.FindElementById("submit").Click();
                    var alertbox = driver.FindElementById("alert");
                    if (alertbox.Text.Contains("disposable"))
                    {
                        return "PASS";
                    }
                    else
                    {
                        return "FAIL: alertbox.Text should contain " + 
                            "the word 'disposable'";
                    }
                }

                catch (Exception ex)
                {
                    return $"FAIL: {ex.Message}";
                }
            }
        }
    }
}

Alternatively you can look at Azure Container Instances with Headless Chrome. There's a .NET SDK as well.