Migrate from PAT to Github App Authentication (GoogleCloudPlatform#773)

Update readme.md.
Migrate from PAT to Github App Authentication.

Author: severus-ho

plugins/README.md

@@ -3,17 +3,21 @@ The "on_failure_actions" plugin listens to the DAGs run results and takes action
 ## Pre-requisites:
 To leverage the "on_failure_actions" plugin, ensure the following conditions are met:
 
-### 1.  **DAG Opt-In:** 
+### 1.  **DAG Opt-In:**
 Each DAG intended to utilize this feature **must include the `"on_failure_alert"` tag** within its DAG definition. DAGs without this specific tag will be ignored by the plugin's failure-handling logic, and no GitHub issue will be filed for their failures.
 
     with DAG(
         dag_id='my_critical_dag',
         # ... other DAG parameters ...
-        tags=['data_ingestion', 'critical', 'on_failure_alert'], # <--- Add this tag
+        tags = [
+                   'data_ingestion',
+                   'critical',
+                   'on_failure_alert' # <--- Add this tag
+               ],
     ) as dag:
         # ... tasks ...
 
-### 2.  **GitHub Owner Mapping:** 
+### 2.  **GitHub Owner Mapping:**
 For accurate issue assignment, ensure that the `owner` property defined for tests within your DAGs corresponds directly to valid **GitHub usernames**. The plugin will collect unique test owners from the failed DAG and attempt to assign the GitHub issue to these users.
 
 #### Example task definition
@@ -29,20 +33,71 @@ For accurate issue assignment, ensure that the `owner` property defined for test
     def task_a():
         pass
 
+### 2.5 **GitHub Owner Mapping for Test Config:**
+If you are using test_config.py in your DAG, you should fill test_owner attribute with your GitHub username
+
+#### Example test definition
+    maxtext_v4_configs_test = gke_config.get_gke_config(
+        num_slices=slice_num,
+        cluster=clusters[accelerator],
+        time_out_in_min=60,
+        test_name=f"maxtext-checkpointing-{mode.value}-{chkpt_mode}",
+        run_model_cmds=command,
+        docker_image=image.value,
+        test_owner='github_username_here',  # This should be a valid GitHub username
+    ).run()
+
 ## Configuration and Installation:
+### Create a GitHub APP
+Refer to this [procedure](https://docs.github.com/en/apps/creating-github-apps/registering-a-github-app/registering-a-github-app) to register a GitHub App.
+
+- Give the app a unique name, for example "ml-auto-solutions-app", and record it as "APP_ID". 
+- Choose "Issues" for the "Repository permissions".
+- Install it to the Target GitHub Repo (ml-auto-solutions). Record the "INSTALLATION_ID" from the web URL. For example, if the URL is https://github.com/apps/ml-auto-solutions-app/installations/78297659, then 78297659 is the INSTALLATION_ID.
+- Create a private key and record it as PRIVATE_KEY.
+
+The APP_ID, INSTALLATION_ID, and PRIVATE_KEY will be used in the next steps.
+
+### Set up via GCP console UI
 1. From GCP console UI, Your Composer Env -> Tab -> Pypi packages -> Edit -> Add 'apache-airflow-providers-github' -> Save
 
-2. From GCP console UI, search for "Secret Manager", and add conn_id 'github_default' into Secret Manager.
-   1. key: airflow-connections-<composer_environment_name>-github_default
-   2. value:
-       {
-           "conn_type": "http",
-           "host": "https://api.github.com",
-           "password": "\<GitHub Personal Access Token\>"
-       }
+2. From GCP console UI, search for "Secret Manager". Go to this page and click '+ CREATE SECRET' button.
+   1. Name: airflow-connections-<composer_environment_name>-github_app
+      Secret value:
+      ```  
+      {
+          "app_id": "<APP_ID>",
+          "installation_id": "<INSTALLATION_ID>",
+          "private_key": "<PRIVATE_KEY>"
+      }
+      ```
+
+3. Upload 'on_failure_actions.py' to \<DAG Bucket\>/plugins/
+
+### Set up via Google Cloud CLI
+1. Set the environment variables in your terminal:
+    ```
+    COMPOSER_ENVIRONMENT_NAME=<Composer Environment Name>
+    COMPOSER_LOCATION=<Composer Location>
+    COMPOSER_DAG_BUCKET=gs://<DAG Bucket>
+    APP_ID=<GitHub App Id>
+    INSTALLATION_ID=<GitHub Installation Id>
+    PRIVATE_KEY=<GitHub App Private Key>
+    ```
+
+2. Add `apache-airflow-providers-github` Pypi package to your Composer environment.
+    ```
+    gcloud composer environments update ${COMPOSER_ENVIRONMENT_NAME} --location=${COMPOSER_LOCATION} --update-pypi-package='apache-airflow-providers-github'
+    ```
+
+3. Add a new Secret into Secret Manager for accessing GitHub App.
+    ```
+    echo "{"app_id": "${APP_ID}", "installation_id": "${INSTALLATION_ID}", "private_key": "${PRIVATE_KEY}"}" | gcloud secrets create airflow-connections-${COMPOSER_ENVIRONMENT_NAME}-github_app --data-file=-
+    ```
+
+4. Upload Composer plugins to the folder for dags in your Composer bucket
+    ```
+    gcloud storage cp ./on_failure_actions.py ${COMPOSER_DAG_BUCKET}/plugins/
+    ```
 
-3. Composer -> Airflow configuration overrides -> Edit, reference: https://cloud.google.com/composer/docs/composer-1/configure-secret-manager
-   1. secrets | backend | airflow.providers.google.cloud.secrets.secret_manager.CloudSecretManagerBackend
-   2. secrets | backends_order | custom,environment_variable,metastore
 
-4. Upload 'on_failure_actions.py' to \<DAG Bucket\>/plugins/