Adobe AD0-E124 Adobe Experience Manager DevOps Engineer Expert Exam Practice Test

In AEM as a Cloud Service, directly deploying packages via the AEM Package Manager UI is severely restricted. You cannot deploy immutable content (code, OSGi configs, or anything under /apps or /libs) via the Package Manager; immutable content must go through the CI/CD pipeline. If a DevOps engineer is forced to use the Package Manager, the package must be strictly limited to mutable content (authorable content, configurations under /conf, /content, /dam). Therefore, they must refactor the package, explicitly set the < packageType > to content, and update the filter.xml to exclude any immutable paths.

When troubleshooting routing or Virtual Host configurations on a Dispatcher (which is typically an Apache or Nginx web proxy), the proxy relies on the Host HTTP header to determine which farm or server block should process the request. Using cURL, you can bypass DNS layers and test the origin server directly by specifying the target IP or default hostname (e.g., origin-mysite.adobeqms.com) while explicitly passing the desired Virtual Host name via the Host header (-H "Host: mysite.com"). This forces the Dispatcher to process the request as if it were traffic coming for mysite.com.

When executing Custom UI Tests (Selenium/Cypress) within an AEM Cloud Manager pipeline, the pipeline looks for a specific Maven module typically named ui.tests. During the dedicated UI Testing phase, Cloud Manager injects an environment variable named CM_BUILD. To ensure the tests are properly triggered and executed inside the Cloud Manager docker container, the Maven profile responsible for the tests (often named ui-tests-docker-execution or similar) must be configured in the pom.xml to activate automatically when the env.CM_BUILD property is present.

In Adobe Cloud Manager, a Sandbox program is specifically designed for training, demos, enablement, and proof-of-concept (PoC) purposes. Unlike Production programs, Sandbox programs typically come pre-loaded with sample content (such as the AEM WKND reference site) and sample configurations to allow users to immediately start exploring the platform. Within a Sandbox program, you can create a Development environment to deploy code and test features. Production programs do not include sample content by default, as they are meant for live, secure, customer-facing applications.

The Dispatcher relies on the /virtualhost section of the farm configuration to route incoming HTTP requests based on the Host header. If a page renders correctly on www.domain.com but returns a blank or default response on domain.com, it strongly indicates that the bare domain (non-www) has not been added to the /virtualhost array for that specific site's farm configuration. The Dispatcher does not know how to handle the request for the missing host and drops or misroutes it.

To verify and analyze slow-loading pages in AEM, the correct tool is "Request Performance" located under the Diagnosis Tools (accessible via the AEM Web Console / Operations Dashboard). This tool tracks incoming HTTP requests to the AEM instance, breaks down the execution time of the request across various internal components and scripts, and allows the DevOps engineer to pinpoint exactly which part of the page rendering is causing the delay. Query Performance (Option A) is specifically for slow JCR searches, not necessarily overall page load times.

Similar to Code Coverage, a Security Rating of 'B' in Cloud Manager usually results in a "Partial Pass" state (3-tier rating system where critical vulnerabilities cause outright failures, but lower-tier warnings cause partial passes). Since all critical and mandatory tests passed, the pipeline pauses and waits for user intervention. The DevOps engineer (with the Deployment Manager role) can proceed with the build by navigating to the execution details, selecting the option to "Override failed results," and submitting the decision to resume the pipeline.

AEM as a Cloud Service enforces a strict separation between mutable content (e.g., /content, /conf) and immutable content (e.g., /apps, /libs). Immutable areas are packaged into the read-only base image during the build process and cannot be modified at runtime. When a Repoinit script (which executes when the repository initializes) attempts to create nodes or set ACLs on paths beneath /libs or /apps, the Jackrabbit Oak repository rejects the change with an UnsupportedOperationException: This builder is read only, subsequently crashing the Sling Repository service and failing the deployment.

To allow an external system (like a CI/CD server or custom dashboard) to programmatically interact with Cloud Manager (such as fetching pipeline metrics), you must set up a server-to-server API integration. This is done via the Adobe Developer Console (Adobe I/O). The DevOps engineer must create a new project, add the Cloud Manager API service, and configure authentication using a public/private key pair (typically generating a JWT token for secure, headless authentication). Adding standard user accounts (Option A) is an anti-pattern for system integrations because it relies on interactive logins or brittle basic auth.

In Adobe Cloud Manager, system-level monitoring metrics and their associated alerting thresholds (like CPU limits, Memory usage, and Segment Store disk space) are managed by Adobe's internal operations team. Standard DevOps engineers and customer administrators do not have the ability to alter these foundational infrastructure thresholds via the Cloud Manager UI or APIs. To request a configuration change to alerting thresholds, the engineer must file a request via their Adobe Customer Success Engineer (CSE) or Adobe Support.

Adobe Support has strict Service Level Agreement (SLA) definitions for priority levels.

P1 (Critical): The production application is completely down and unavailable to end-users (e.g., Publish tier is crashed).

P2 (Serious): A core, business-critical function is down for all users, causing severe revenue or operational impact.

P3 (Interruption to normal operations): The application is functional, but specific users are experiencing degraded performance or feature loss (like a broken authoring workflow).

Even though the business has a deadline (a product launch within 24 hours), the fact that the live production site is still up and only two authors are currently reporting an inability to activate pages categorizes this strictly as a P3 issue according to standard Adobe support guidelines.

If a Cloud Manager build requires dependencies from a private, password-protected corporate artifact repository, the DevOps engineer must provide Cloud Manager with a custom Maven settings file to handle the authentication. This is done by placing a file at .cloudmanager/maven/settings.xml in the root of the AEM Git repository. Because hardcoding Basic Authentication credentials in Git is a massive security risk, the credentials inside the settings.xml should reference environment variables. The DevOps engineer then uses the Adobe I/O CLI (aio cloudmanager:set-pipeline-variables) to inject the actual username and password as secure pipeline secrets.

When AEM search results are outdated and the LastIndexedTime metric is stalled, it indicates that the Oak Asynchronous Indexer has failed or is stuck on a corrupted node. To diagnose Oak indexing issues, you must create a specific logger in AEM OSGi configurations for the Java package responsible for Oak indexing: org.apache.jackrabbit.oak.plugins.index. Setting this logger to DEBUG or TRACE will output detailed information into the logs, revealing the exact path or query causing the indexer to fail.

Intermittent 503 Service Unavailable errors on publish instances during peak traffic, specifically tied to heavy image rendition generation, indicate that the Publish servers' CPUs are being overwhelmed by repetitive processing. In AEM, these dynamic renditions should only be generated once and then cached by the Dispatcher. If the load is hitting the publish tier constantly, it means the Dispatcher is not caching those rendition requests. The DevOps engineer must check the /rules section under /cache in the dispatcher farm configuration to ensure that the URL patterns for these image renditions are permitted to be cached.

To manually migrate a package between two AEM instances via the command line, the process involves two distinct curl commands using the Jackrabbit Package Manager API.

Download: You must download the generated .zip file from the source (STG) environment using -O to save the remote file locally.

Upload & Install: You must upload the local file to the target (PROD) environment and simultaneously install it by passing the -F file=@ < filename > along with the multipart form parameters -F install=true and -F force=true to the /crx/packmgr/service.jsp endpoint.