Start testing with the gcloud CLI

This guide describes how to run an instrumentation, Robo, or Game Loop test using the gcloud CLI.

For a complete list of gcloud commands you can use with your Android app in Test Lab, visit the reference documentation for gcloud firebase test android.

Before you begin

If you haven't already, add Firebase to your Android project.

Step 1. Set up the gcloud CLI

1. Download the Google Cloud SDK

This includes the gcloud CLI tool.

2. Make sure your installation is up-to-date:

   gcloud components update
   

3. Sign in to the gcloud CLI using your Google Account:

   gcloud auth login
   

4. Set your Firebase project in gcloud, where PROJECT_ID is the ID of your Firebase project:

   gcloud config set project PROJECT_ID
   

Step 2. Check available test devices

Use the following gcloud commands to view test devices and locales that are available for your test.

As an option, you can also download the sample Notepad app to start running the commands right away. Use the binary file app-debug-unaligned.apk and instrumentation tests file app-debug-test-unaligned.apk, which are located in NotePad/app/build/outputs/apk/.

• models list: Get a current list of Android devices available for you to test against.

  gcloud firebase test android models list

  In the command output:

  • Column MODEL_ID contains the identifier you can later use to run tests on the device model.
  • Column OS_VERSION_ID contains the operating system versions supported by the device.

  Example output

  

• models describe: Get more information about a specific Android MODEL_ID.

  gcloud firebase test android models describe MODEL_ID

  The output contains the device model's brand, manufacturer, OS versions, supported API levels, supported Application Binary Interfaces (ABI), release dates, and whether the device is physical or virtual.

• versions list: Get a list of currently available OS versions to test against.

  gcloud firebase test android versions list

  You can use an identifier from either of the first two columns of command output (OS_VERSION_ID and VERSION), to later run tests against an Android OS version. If you don't specify the Android OS versions to test against, the default noted under the TAGS column is used.

  Example output

  

• locales list: Get the current list of locales available to test against.

  gcloud firebase test android versions list

  The first column of the command output, LOCALE, contains the identifier that you can use later to run tests against a locale. If you don't specify the locales to test against, English is used as the default locale.

Step 3. Run your test

Now that you know the range of device models, locales, and OS versions available for testing your app, you can specify devices using the gcloud firebase test android run command and the --device flag to run Robo or instrumentation tests.

Note: The --device flag is now the preferred way to specify test devices and may not be used in conjunction with --devices-ids, --os-version-ids, --locales, or --orientations. To learn about these legacy flags, see the list of deprecated device dimension flags at gcloud firebase test android run.

Run a Robo test

Even if you don't have any instrumentation tests, you can still look for bugs in your app. Use the Robo test to perform an automated review of your app's user interface. Robo test exercises the app by performing a static analysis of the various paths through the app's user interface, and then crawls through the app to find crashes and other potential issues.

To run a Robo test, run the following example command:

gcloud firebase test android run \
  --type robo \
  --app app-debug-unaligned.apk \
  --device model=Nexus6,version=21,locale=en,orientation=portrait  \
  --device model=Nexus7,version=19,locale=fr,orientation=landscape \
  --timeout 90s \
  --client-details matrixLabel="Example matrix label"

• The --type robo parameter is implicit if no --type value is specified.
• To help you identify and locate your test matrices in the Firebase console, you can use the optional --client-details matrixLabel="Example matrix label" flag to label your test matrix.
• You can see the complete set of command line options for running tests by typing: gcloud help firebase test android run.

As an alternative to specifying these arguments on the command line, you can optionally specify your arguments in a YAML-formatted argument file. Run gcloud topic arg-files to learn how to use this feature.

To learn how to investigate the test results from the Robo test, See Analyze your test results.

Run an instrumentation test

Now use the gcloud command line tool to run the Notepad app's Espresso tests on your specified Android device configurations. Use the instrumentation test type to run the tests in app-debug-test-unaligned.apk as follows:

gcloud firebase test android run \
  --type instrumentation \
  --app app-debug-unaligned.apk \
  --test app-debug-test-unaligned.apk \
  --device model=Nexus6,version=21,locale=en,orientation=portrait  \
  --device model=Nexus7,version=19,locale=fr,orientation=landscape
  --client-details matrixLabel="Example matrix label"

• The --type instrumentation parameter is implicit if a test APK is specified with --test.
• To help you identify and locate your test matrices in the Firebase console, you can use the optional --client-details matrixLabel="Example matrix label" flag to label your test matrix.
• You can see the complete set of command line options for running tests by typing gcloud help firebase test android run.

As an alternative to specifying these arguments on the command line, you can optionally specify your arguments in a YAML-formatted argument file. Run gcloud topic arg-files to learn how to use this feature.

The gcloud CLI supports Android Test Orchestrator. Orchestrator requires AndroidJUnitRunner v1.1 or higher. To enable it, use gcloud firebase test android run with the
--use-orchestrator flag. To disable it use the --no-use-orchestrator flag.

Caution: To avoid spending quota on or being billed for malfunctioning tests, run Orchestrator locally before trying it in Test Lab. Confirm that Orchestrator is working in your app by running a test on your own machine before uploading your APK. Keep in mind that testing with Orchestrator takes slightly longer than without it, and it might impact your billing or cause your tests to exceed your timeout limit.

You can also control how Test Lab runs your instrumentation tests using additional flags that are not shown above. For example, you can use the --test-targets flag to test a single class or a class method used by your test APK. You can also find out whether your test that failed was actually flaky or not by using --num-flaky-test-attempts flag, which specifies the number of times a test execution should be re-attempted if one or more of its test cases fail for any reason. To learn more, see gcloud firebase test android run.

Code coverage reports for instrumentation tests

Test Lab supports code coverage reporting tools EMMA and JaCoCo. If you have either tool integrated into the build for your app, you can get a code coverage report for Test Lab tests by running gcloud firebase test android run with some additional arguments. If Android Test Orchestrator is not enabled, use the following:

gcloud firebase test android run \
  --type instrumentation \
  --app your-app.apk \
  --test your-app-test.apk \
  --device model=TestDevice,version=AndroidVersion  \
  --environment-variables coverage=true,coverageFile="/sdcard/Download/coverage.ec" \
  --directories-to-pull /sdcard/Download

If you are generating code coverage reports while also using Android Test Orchestrator, modify your environment variables as follows:

gcloud firebase test android run \
  --type instrumentation \
  --app your-app.apk \
  --test your-app-test.apk \
  --device model=TestDevice,version=AndroidVersion  \
  --environment-variables clearPackageData=true,coverage=true,coverageFilePath="/sdcard/Download/" \
  --directories-to-pull /sdcard/Download

When Test Lab finishes running your tests, find your code coverage reports in Google Cloud Storage:

1. Open the Firebase console link that the gcloud tool printed above the test result table in your terminal.
2. Click a test execution from the list at that link to open that execution's details page.
3. Click Test results to go to the Cloud Storage bucket with that execution's test results.
4. Open artifacts/coverage.ec to see your code coverage report.

Note: Pulling the sdcard/Download directory for each test places all files from that directory into the test's Cloud Storage bucket. If you're on a Blaze payment plan and your app creates many files in that directory, pulling the directory might result in extra storage charges.

Analyze your test results

After a few minutes, a basic summary of your test results is printed by the gcloud tool:

The output of your command line test run also includes a link to view test results. To learn more about how to interpret these results, see Analyzing Firebase Test Lab for Android Results.

Custom login and text input with Robo test

Robo test automatically completes sign-in screens that use a Google account for authentication, unless you use the --no-auto-google-login parameter. It can also complete custom login screens using test account credentials that you provide. You can also use this parameter to provide custom input text for other text fields used by your app.

To complete text fields in your app, use the --robo-directives parameter and provide a comma-separated list of key-value pairs, where the key is the Android resource name of the target UI element, and the value is the text string. You can also use this flag to tell Robo to ignore specific UI elements (e.g., "logout" button). EditText fields are supported but not text fields in WebView UI elements.

For example, you could use the following parameter for custom login:

--robo-directives username_resource=username,password_resource=password

Caution: Only use credentials for test accounts that are not associated with real users.

Available commands and flags

The Test Lab gcloud CLI has several commands and flags available that let you run tests with different specifications:

• Android Test Orchestrator flag: A flag to enable Orchestrator, a tool that allows you to run each of your app's tests in its own invocation of Instrumentation. Test Lab always runs the latest version of Orchestrator.

• Game Loop test flags: A set of config flags that enable and control a "demo mode" to simulate player actions in game apps. Learn more about running Game Loop tests with Test Lab.

• Uniform Sharding flag (in beta): A flag that specifies the number of shards into which you want to evenly distribute test cases. The shards are run in parallel on separate devices.

  Note: The limits for physical devices are 50, 200 for Arm virtual, and 500 for x86 virtual. This also applies to the Manual Sharding flag.

• Manual Sharding flag (in beta): A flag that specifies a group of packages, classes, and/or test cases to run in a shard (a group of test cases). The shards are run in parallel on separate devices.

• Network traffic profiles flag (in beta): A flag that specifies which network profile your tests use with physical devices. Network profiles emulate a variety of networks conditions, allowing you to test your app's performance on unreliable or unpredictable networks.

Scripting gcloud commands with Test Lab

You can use shell scripts or batch files to automate mobile app testing commands that you would otherwise run using the gcloud command line. The following example bash script runs an instrumentation test with a two-minute timeout, and reports if the test run completed successfully:

if gcloud firebase test android run --app my-app.apk --test my-test.apk --timeout 2m
then
    echo "Test matrix successfully finished"
else
    echo "Test matrix exited abnormally with non-zero exit code: " $?
fi

Script exit codes

Test Lab provides several exit codes that you can use to better understand the results of tests that you run using scripts or batch files.

Scripting exit codes for Test Lab

Exit code	Notes
0	All test executions passed.
1	A general failure occurred. Possible causes include: a filename that does not exist or an HTTP/network error.
2	Testing exited because unknown commands or arguments were provided.
10	One or more test cases (tested classes or class methods) within a test execution did not pass.
15	Firebase Test Lab could not determine if the test matrix passed or failed, because of an unexpected error.
18	The test environment for this test execution is not supported because of incompatible test dimensions. This error might occur if the selected Android API level is not supported by the selected device type.
19	The test matrix was canceled by the user.
20	A test infrastructure error occurred.