Salesforce Development Lifecycle and Deployment Architect Practice Questions with Explanations
Free Salesforce Development Lifecycle and Deployment Architect practice questions. 46 of them, each with the correct answer, a full explanation, and the reason every other option is wrong.
Development Lifecycle and Deployment Architect practice questions
Question 1
Northern Trail Outfitters (NTO) recently acquired Eastern Trail Outfitters (ETO). NTO’s sales leadership team had hands-on experience with the ETO’s Sales Optimization app and have given the feedback that the app would benefit NTO’s sales team.
Which option should the architect recommend for having ETO’s Sales Optimization app in NTO’s Salesforce org in the shortest possible time?
- A. Create users in ETO’s org and provide access to NTO’s sales team.
- B. Create a core team and build the Sales Optimization app in NTO’s org.
- C. Create a managed package of the app and deploy in NTO’s org.
- D. Create an unmanaged package of the app and deploy in NTO’s org.
Show answer and explanation ▾
Correct answer: D
An unmanaged package copies the app's components into NTO's org with no rebuild and no user setup, so it is the quickest way to get the app there.
Why the other options are wrong:
- A. Creating users in ETO's org leaves the app in a separate org instead of NTO's org.
- B. Rebuilding the app in NTO's org takes far longer than moving the existing one.
- C. A managed package needs a packaging org and a namespace and is meant for distribution, so it takes longer here.
Question 2
Which are two recommended methods of creating test data in Salesforce? (Choose two.)
- A. Host a mock endpoint to produce sample information from an endpoint.
- B. Reference data from middleware directly within your test class.
- C. Utilize Heroku Connect to provide test class data.
- D. Load a CSV as a static resource and reference it in a test class.
Show answer and explanation ▾
Correct answer: A, D
Good test data comes from sources a test can control: a mock endpoint that returns sample information, and a CSV loaded as a static resource that the test class reads.
Why the other options are wrong:
- B. Referencing middleware data directly from a test class makes tests depend on an outside system.
- C. Heroku Connect syncs live data and is not a way to supply test class data.
Question 3
Universal Containers (UC) has recently acquired other companies that have their own Salesforce orgs. These companies have been merged as new UC business units.
The CEO has requested an architect to review the org strategy. taking into consideration two main factors:
The CEO wants business process standardization among all business units.
Business process integration is not required as the different business units have different customers and expertise.
Which org strategy should the architect recommend in this scenario, and why?
- A. Single-org strategy, as costs increase as the number of orgs go up.
- B. Multi-org strategy, as it is uncommon for the diversified business units to get used to working in the same space as the other business units.
- C. Multi-org strategy, as they could deploy a common managed package into the orgs of the different business units.
- D. Single-org strategy, as the high level of business process standardization will be easier to implement in a single org.
Show answer and explanation ▾
Correct answer: D
When the main aim is standardized business processes and integration is not needed, a single org makes standardization easier to implement.
Why the other options are wrong:
- A. Cost is not the reason: the stated driver is process standardization.
- B. Whether business units like sharing a space is not what the CEO asked the architect to weigh.
- C. A common managed package in several orgs does not deliver the same level of standardization as one org.
Question 4
Universal Containers (UC) is planning to move to Salesforce Sales Cloud and retire its homegrown on-premise system. As part of the project, UC will need to migrate 5 million Accounts, 10 million Contacts, and 5 million Leads to Salesforce.
Which three areas should be tested as part of data migration? (Choose three.)
- A. Account and Lead ownership
- B. Contact association with correct Account
- C. Page layout assignments
- D. Lead assignment
- E. Data transformation against source system
Show answer and explanation ▾
Correct answer: A, B, E
Migration testing checks that records kept the right owner, that contacts link to the correct accounts, and that the transformed data matches the source system.
Why the other options are wrong:
- C. Page layout assignments are configuration, not migrated data.
- D. Lead assignment is rule behavior, not data that was migrated.
Question 5
Universal Containers (UC) is a high-tech company using SFDX tools and methodologies for its Salesforce development. T UC has moved some of its code and configuration to Unlocked Packages.
Which two best practices should an architect recommend to support UC’s new package development strategy? (Choose two.)
- A. Move everything in the existing codebase to a single monolithic package.
- B. Version control does not need to be used, as packages manage all the code and configuration.
- C. Test developed packages in test environments before installing to production.
- D. Consult the metadata coverage report to identify features supported by packages.
Show answer and explanation ▾
Correct answer: C, D
Packages should be tested in test environments before production, and the metadata coverage report shows which features packages support.
Why the other options are wrong:
- A. A single monolithic package defeats the modular design that packaging is meant to give.
- B. Version control is still needed as the source of truth, because packages do not replace it.
Question 6
Universal Containers (UC) has been on the org development model with scratch orgs are already enabled, but they haven’t been taking advantage of the scratch orgs.
Now UC is ready to move to the package development model.
What step must be done by an administrator?
- A. In setup, switch the Enable Unlocked Packages to Enabled, keep the Enable Second-Generation Managed Packages as disabled.
- B. In setup, switch the Enable Dev Hub to Enabled, then switch the Enable Source Tracking for Scratch Orgs to Enabled.
- C. In setup, switch the Enable Unlocked Packages and Second-Generation Managed Packages to Enabled.
- D. In setup, switch both the Enable Dev Hub and Enable 2nd-Generation Managed Packages to Enabled.
Show answer and explanation ▾
Correct answer: C
The package development model relies on second-generation packaging, which an administrator turns on in the Dev Hub org with the setting Enable Unlocked Packages and Second-Generation Managed Packages. Dev Hub is already on because scratch orgs are enabled, so this is the remaining step.
Why the other options are wrong:
- A. Unlocked packages and second-generation managed packages are enabled together by one setting, so they cannot be split this way.
- B. Dev Hub is already enabled, and source tracking does not turn on packaging.
- D. Dev Hub is already enabled in this org, so enabling it is not the missing step.
Question 7
Universal Containers (UC) has multi-cloud architecture in a single org. The Sales Cloud dev team is working in a Dev Pro sandbox (DevPro1) of delivering a release in three months. The business requirements from Service Cloud warrant a quicker release in four weeks, but need part of the Sales Cloud work that is completed in DevPro1. The decision of using a separate Dey Pro sandbox (DevPro2) is still pending. The DevPro1 was upgraded to preview for next salesforce major release two weeks ago.
What should an Architect recommend?
- A. Clone the DevPro1 sandbox and name it DevPro2 for the second work stream to work on the Service Cloud requirements.
- B. Push back on the requirements because adding another work stream will bring some risks with it.
- C. Ask the second work stream team to work on the same DevPro1 sandbox.
- D. DevPro1 cannot be cloned because it is on a different version from Production. Just create a new DevPro2, and migrate metadata from DevPro1.
Show answer and explanation ▾
Correct answer: A
Cloning DevPro1 gives the second work stream its own sandbox that already contains the completed Sales Cloud work, so the Service Cloud team can build on it and release on a separate four-week timeline without disturbing the first team.
Why the other options are wrong:
- B. Refusing a valid business requirement is not an architectural recommendation when the platform supports parallel work streams.
- C. Sharing one sandbox mixes two releases with different dates and makes it hard to deploy only the Service Cloud work.
- D. Creating an empty sandbox and migrating the metadata by hand is slower and more error prone than cloning, and it is not required.
Question 8
Universal Containers (UC) Customer Community is scheduled to go live in the Europe, Middle East, and Africa (EMEA) region in 3 months. UC follows a typical centralized governance model. Two weeks ago, the project stakeholders informed the project team about the recent changes in mandatory compliance requirements needed to go live. The project team analyzed the requirements and have estimated additional budget needs of 30% of the project cost for incorporating the compliance requirements.
Which management team is empowered to approve this additional budget requirements?
- A. Security Review Committee
- B. Executive Steering Committee
- C. Project Management Committee
- D. Change Control Board
Show answer and explanation ▾
Correct answer: B
In a centralized governance model the executive steering committee owns the program's funding and strategic direction, so a 30% budget increase driven by new compliance requirements goes to that group for approval.
Why the other options are wrong:
- A. A security review committee assesses security risk and does not approve project budgets.
- C. A project management committee runs day-to-day delivery and escalates budget increases of this size.
- D. A change control board approves changes to scope and releases, not additional program funding.
Question 9
Universal Containers uses multiple Salesforce orgs for its different lines of business (LOBs). In a recent analysis, the architect found that UC could have a more complete view of its customers by gathering customer data from different orgs.
What two options can an architect recommend to accomplish the customer 380-degree view? (Choose two.)
- A. Implement a Complete Graph multi-org strategy by allowing each org to connect directly to every other, reading and writing customer data from the orgs where it has been originally created.
- B. Migrate from multi-org to single-org strategy, consolidating customer data in the process.
- C. Implement a Single Package multi-org strategy by developing and deploying to all orgs a managed package which reads and consolidates customer 360-degree view from the different orgs.
- D. Implement a Hub-and-Spoke multi-org strategy by consolidating customer data in a single org, which will be the master of customer data, and using integration strategies to let the LOBs orgs read and write from it.
Show answer and explanation ▾
Correct answer: B, D
A complete customer view needs one place where customer data is mastered. Consolidating into a single org achieves that directly, and a hub-and-spoke design achieves it by making one org the customer master that the line-of-business orgs read from and write to through integration.
Why the other options are wrong:
- A. Connecting every org directly to every other creates many point-to-point integrations and leaves no single master of customer data.
- C. Deploying the same managed package to every org does not consolidate customer data into one master record.
Question 10
Universal Containers (UC) is considering implementing a minor change policy for a series of low-risk user stores that are commonly received by the UC admins. The policy would allow admins to make these changes directly in production. UC does not have continuous integration/continuous delivery (CI/CD) in place.
Which three best practices should the architect suggest UC follow for their new change policy? (Choose three.)
- A. All changes should still be tested.
- B. CI/CD is required in to successful manage minor changes.
- C. Downstream environments will not be automatically updated when production changes.
- D. Minor changes do not need to be documented and can be made at any time.
- E. Minor changes should be thoroughly documented and follow some type of standard cadence.
Show answer and explanation ▾
Correct answer: A, C, E
Low-risk changes made directly in production still need testing, documentation and a regular cadence so they stay controlled and auditable. Because sandboxes are not updated when production changes, the team must also plan to bring those changes back into downstream environments.
Why the other options are wrong:
- B. Minor changes can be managed with a clear manual process, so CI/CD is not a requirement.
- D. Undocumented changes made at any time remove traceability and cause conflicts with later deployments.
Question 11
Universal Containers is adopting Scrum as an agile methodology and wants to choose a software tool to support the adoption.
What three key features of an agile development support tool should an architect look for? (Choose three.)
- A. Work (for example, user stories or tasks) assignment
- B. Email notifications when work is created or changed
- C. Sprint backlog management
- D. Kanban board
- E. Product backlog prioritization
Show answer and explanation ▾
Correct answer: A, D, E
A tool that supports Scrum must let the team assign user stories and tasks, show the flow of work on a board, and let the product owner order the product backlog by priority, since these drive daily work and sprint planning.
Why the other options are wrong:
- B. Email notifications are a convenience and not a core capability for running Scrum.
- C. Managing the sprint backlog follows from work assignment and the board, so it is not a separate selection criterion here.
Question 12
Universal Containers (UC) is looking for advice on how often it should refresh its sandboxes. UC currently uses a development lifecycle that starts with developer environments and moves to integration testing, QA testing, UAT, and then production. They have many scrum teams working concurrently and the teams do not agree on when refreshes should occur.
What two recommendations should the architect suggest? (Choose two.)
- A. Sandboxes should be refreshed on the day when the refresh is allowed for that type of sandbox.
- B. Development environments should generally be refreshed after each working feature has been successfully migrated.
- C. Production is the only pristine environment.
- D. Integration sandboxes should be refreshed rarely because of the burden of maintaining the various API.
Show answer and explanation ▾
Correct answer: B, C
Production is the only environment guaranteed to hold the true state of the org, so sandboxes drift from it over time. Refreshing a development environment once its feature has been migrated brings it back in line with production before new work starts.
Why the other options are wrong:
- A. Refreshing as soon as the interval allows ignores work in progress and can wipe out unfinished changes.
- D. Integration sandboxes need regular refreshes to stay aligned with production, so rare refreshes are not the recommendation.
Question 13
Universal Containers has seven orgs in different regions. Its processes are global and standardized but each region needs the flexibility to be able to understand the global code and customize some aspects for its regions.
Which development model is optimized for this need?
- A. Use unlocked packages to deploy the global code and allow each country to create its customized unlocked package extensions.
- B. Use a managed package to deploy the global code and allow local teams to request the addition of code within that package.
- C. Create a centralized Git with all the code and where the global team approves the changes made by the local teams.
- D. Use a managed package for global code and another managed package for all regions code.
Show answer and explanation ▾
Correct answer: A
Unlocked packages keep their source visible in the org where they are installed, so regional teams can read the global code. Each region can then build its own unlocked package that depends on and extends the global package.
Why the other options are wrong:
- B. A managed package hides its code from subscribers and forces every regional change through the central team.
- C. A single repository with central approval does not give regions independently deployable customizations.
- D. Managed packages obscure the code, so regions could not understand the global logic, and one shared regional package removes local flexibility.
Question 14
Universal Containers (UC) is on the Unlimited Edition of Salesforce orgs. Currently there are four Developer Pro sandboxes used by dev teams for various purposes, one partial copy sandboxes used for training, one full sandbox used for user acceptance testing. There is another full sandbox used for performance testing during the initial launch, and it is no longer being actively used.
A system administrator had reported yesterday that none of the sandboxes can be refreshed.
What is the most probable action an architect can help the system administrator with?
- A. Create a Salesforce support case, someone should know what is wrong.
- B. Delete the full sandbox used for performance testing.
- C. It appears the org is using more sandboxes than the license permits. The Partial Copy sandbox shouldn’t be used.
- D. Contact the Salesforce Account Team to do a sandbox License Count.
Show answer and explanation ▾
Correct answer: B
Each sandbox type is limited by the licenses the org holds, and an unused sandbox still consumes one. Deleting the Full sandbox that is no longer needed for performance testing frees a license and is the practical first action.
Why the other options are wrong:
- A. Opening a support case without a specific finding is not the most probable fix for a license allocation problem.
- C. Unlimited Edition includes a Partial Copy sandbox, so the training sandbox is permitted.
- D. The sandbox license counts are already visible in Setup, so the account team is not needed to count them.
Question 15
Northern Trail Outfitters (NTO) has well-defined release management processes for both large and small projects. NTO’s development team created a workflow and a trigger for the changes in its opportunity renewal process.
What should the architect recommend for release planning of these changes?
- A. Plan this as a minor release with training and change management.
- B. Plan this as a major release and align with a Salesforce major release.
- C. Plan this as a patch release and align with the Salesforce patch release.
- D. Plan this an interim release after checking with Salesforce support.
Show answer and explanation ▾
Correct answer: A
A workflow and a trigger that change one business process are a small, contained set of changes. That fits a minor release, which still includes user training and change management because the renewal process behaves differently afterward.
Why the other options are wrong:
- B. A major release is for large sets of new functionality and does not need to follow the Salesforce release schedule.
- C. A patch release is for bug fixes, not for new automation that changes a process.
- D. Salesforce support does not need to be consulted to schedule a customer's own release.
Question 16
Sales and Service products will be created by two teams that will use second-generation managed package(s). The Sales team will use a specific function of the Service product, but the architect wants to ensure that this team will only use the functions exposed by the Service team. No other team will use these same functions.
What should an architect recommend?
- A. Create two second generation managed packages with the same namespace and set the methods that should be shared with the @namespaceAccessible annotation.
- B. Create two managed packages with Sales and Service namespaces. Set the methods to be shared with the @salesAccessible annotation.
- C. Create two managed packages. Create an authentication function in the Service package that will return a token if a Sales user is authorized to call the exposed function. Validate the token in the Service functions.
- D. Create a managed package with both products and create a code review process with an approver from each team.
Show answer and explanation ▾
Correct answer: A
Second-generation managed packages that share a namespace can call each other's Apex when it is marked with the @namespaceAccessible annotation. The Service team annotates only the methods it wants to expose, and code outside that namespace cannot use them.
Why the other options are wrong:
- B. There is no @salesAccessible annotation in Apex, and packages in different namespaces cannot share non-global code this way.
- C. A custom token scheme adds complexity and is not the platform mechanism for sharing Apex between packages.
- D. A single package with a review process does not technically restrict which Service functions the Sales code can call.
Question 17
What are three advantages of the package development model? (Choose three.)
- A. Eliminating the need of using change set, which should no longer be used as it can get messy working with package development models.
- B. Providing its own source control, so the source can be deployed in any sandbox orgs.
- C. Improving team development and collaboration.
- D. Facilitating automated testing and continuous integration.
- E. Significantly reducing the need for manually tracking changes.
Show answer and explanation ▾
Correct answer: C, D, E
In the package development model the source of truth is version control and changes are organized into versioned packages. Teams can work in parallel, pipelines can build and test packages automatically, and source tracking replaces manual lists of changed components.
Why the other options are wrong:
- A. Change sets remain available, and removing them is not an advantage of the model.
- B. Packages do not provide source control: the model depends on an external version control system.
Question 18
Universal Containers has started building a customer Lightning community that contains a few dozen Aura components. The development team lead has come to the Salesforce architect about questions regarding testing the Lightning components.
What two knowledge points can the architect pass to the development team lead? (Choose two.)
- A. The testing can be viewed in the lightning.force.com/c/jasminetests.app URL. The page loads and runs Jasmine test and writes pass/fail information to the screen.
- B. There is a $T test helper object that can be used to create the instance of the Lightning component, and its promise enabled.
- C. Install the Lightning test service AppExchange package to enable the Aura component testing.
- D. The testing of the JavaScript part of the Aura component can be tested in the Jest framework and the Apex controllers can be tested using test classes.
Show answer and explanation ▾
Correct answer: A, B
The Lightning Testing Service runs Jasmine tests for Aura components inside a test app in the org and shows pass or fail results on the page. Its $T helper object creates component instances and returns promises, which simplifies asynchronous test code.
Why the other options are wrong:
- C. Installing an AppExchange package is not one of the two points about how these component tests are run and written.
- D. Jest is the framework for Lightning web components, not for the JavaScript in Aura components.
Question 19
Universal Containers (UC) has created a custom REST web service. This web service receives Orders and Order Line Items data from an external endpoint and runs business logic and validations on it before inserting it into the database. UC is expecting to receive more than 100K orders a day and each order can have up to 10 line items. Each inbound request will contain only one order and its corresponding line items.
What two testing types should an architect recommend to ensure users don’t face platform slowdowns during peak business hours? (Choose two.)
- A. Stress Testing
- B. Load Testing
- C. Unit Testing
- D. Performance Testing
Show answer and explanation ▾
Correct answer: A, B
Load testing confirms the service handles the expected volume of more than 100,000 order requests a day at peak, and stress testing pushes beyond that volume to find the point where the service degrades. Together they show whether users will see slowdowns.
Why the other options are wrong:
- C. Unit testing checks individual pieces of logic and says nothing about behavior under heavy request volume.
- D. Performance testing as a general label does not name the specific volume tests this scenario calls for.
Question 20
All AppExchange products are subject to Salesforce security reviews.
What is the most common reason that the prospect AppExchange products fail the security review?
- A. SOQL injection
- B. Cross-site scripting
- C. CRUD/FLS (field level security)
- D. Session hacking
Show answer and explanation ▾
Correct answer: C
Apex runs in system context by default, so code must check object permissions and field-level security itself. Missing CRUD and FLS enforcement is the most frequent finding that causes a package to fail the AppExchange security review.
Why the other options are wrong:
- A. SOQL injection is checked in the review but is not the most frequent cause of failure.
- B. Cross-site scripting is checked in the review but is not the most frequent cause of failure.
- D. Session hacking is not the most frequent finding in the security review.
Question 21
Universal Containers is having trouble aligning releases between major, minor, and Salesforce seasonal releases.
What should an architect recommend?
- A. Share the test plans between the teams on each release type.
- B. Gate all release decisions at the center of excellence.
- C. Create a spreadsheet of metadata changes and reconcile the overlaps.
- D. Create a release calendar, train and align all the teams.
Show answer and explanation ▾
Correct answer: D
A shared release calendar shows major releases, minor releases and the three Salesforce seasonal releases on one timeline. Training every team on it lets them plan work, sandbox refreshes and freezes around the same dates.
Why the other options are wrong:
- A. Sharing test plans improves testing but does not coordinate release dates.
- B. Routing every release decision through the center of excellence creates a bottleneck and does not align the schedules.
- C. A spreadsheet of metadata changes helps find conflicts but does not align release timing.
Question 22
A Salesforce partner intends to build a commercially available application by creating a managed package for distribution through AppExchange.
What two types of environments can the partner use for development of the managed package? (Choose two.)
- A. Developer sandbox
- B. Developer Pro sandbox.
- C. Developer Edition
- D. Partner Developer Edition
Show answer and explanation ▾
Correct answer: C, D
A managed package is developed in an org that can hold the package's namespace. Developer Edition and Partner Developer Edition orgs both support registering a namespace and building a managed package for AppExchange distribution.
Why the other options are wrong:
- A. A Developer sandbox is a copy of a production org and cannot be the packaging org for a managed package.
- B. A Developer Pro sandbox is a copy of a production org and cannot be the packaging org for a managed package.
Question 23
Which two statements are accurate about why Mock objects are needed when writing test classes? (Choose two.)
- A. A Mock is needed whenever the code makes an HTTP callout.
- B. Some methods are invoking long running processes, using Mock is a shortcut of bypassing the long executions.
- C. Mock can also be used on the classes that extend the batchable interface to bypass the batch jobs.
- D. Using a Mock allows the test class to bypass the dependencies of other objects, methods, state, or behaviors. Therefore, the developer has total control of his own code.
Show answer and explanation ▾
Correct answer: A, D
Apex tests cannot make real HTTP callouts, so code that calls an external service must be tested with a mock response. More generally, a mock replaces a dependency with a controlled stand-in, so the test isolates the code under test and its results are predictable.
Why the other options are wrong:
- B. Mocks exist to isolate dependencies, not to shorten long-running processes.
- C. Batch Apex is tested by running the job between Test.startTest and Test.stopTest, not by mocking the batchable class.
Question 24
Universal Containers is about to begin the release of a major project. To facilitate this, they have several sandboxes to make their deployment train. These sandboxes are a mix of preview and non-preview instances.
What should the architect recommend?
- A. No advice needed, mixing instance types is important for regression testing.
- B. Refresh all non-preview sandboxes when the release management team has time.
- C. Contact support to rollback the release when Salesforce upgrades the sandboxes.
- D. Refresh all non-preview sandboxes during the release preview window.
Show answer and explanation ▾
Correct answer: D
A deployment path that mixes preview and non-preview sandboxes can fail because the environments run different platform versions. Refreshing the non-preview sandboxes during the sandbox preview window moves them to preview instances, so every environment in the path is on the same release.
Why the other options are wrong:
- A. Mixed versions in one deployment path cause deployment errors and are not a regression testing practice.
- B. The refresh must happen inside the preview window, not whenever the team has time.
- C. Salesforce support does not roll back a seasonal release on sandboxes.
Question 25
What are two limitations an architect should consider when designing a strategy for managing technical reference data, with multiple related objects? (Choose two.)
- A. HTTP response size
- B. Circular relationships
- C. Depth of nested relationships
- D. Apex CPU limits
Show answer and explanation ▾
Correct answer: B, C
Reference data that spans several related objects must be loaded parent first, so records that refer to each other in a circle cannot be loaded in one pass. Tools that move related records together also limit how many levels of nested relationships one operation can carry.
Why the other options are wrong:
- A. HTTP response size is not the constraint that shapes how related reference records are moved between orgs.
- D. Apex CPU limits apply to code execution and are not the constraint on moving related reference data.
Question 26
Salesforce has three major releases a year.
Which type of change introduced by a release can cause automated browser tests to need updating?
- A. Metadata schema changes
- B. New Apex methods
- C. DOM changes
- D. New standard fields
Show answer and explanation ▾
Correct answer: C
Automated browser tests find buttons, fields and other elements through the page structure. When a Salesforce release changes the DOM of standard pages, the locators in those tests can stop matching and need to be updated.
Why the other options are wrong:
- A. Metadata schema changes affect deployments and API calls, not how a browser test locates page elements.
- B. New Apex methods are additions that existing browser tests do not depend on.
- D. New standard fields do not change how existing tests locate the elements they already use.
Question 27
Universal Containers (UC) has developed extensions of Salesforce Service Cloud for the use of its customer service teams using the change set development model.
Recently, UC acquired a company that develops extensions of an AppExchange app. The development team of the acquired company uses the org development model. The Universal Containers CTO wants both teams to work on a single org and follow the same set of processes.
Which development model should the architect recommend to be used by the consolidated development team?
- A. Package development model, so teams can build release artifacts that can be tested and released independently from artifacts for other projects.
- B. Org development model, because the acquired company’s team is already using it, and it is better than the change set development model.
- C. Change set development model, because UC is already using it, so it will face less resistance.
- D. Package development model, because it allows packages to be created and deployed using declarative (point-and-click) development tools, without writing code.
Show answer and explanation ▾
Correct answer: A
With the package development model, each team's work becomes a versioned package that is built, tested and released on its own schedule. Two teams with different projects can share one org and one process without their releases depending on each other.
Why the other options are wrong:
- B. The org development model still treats the org's metadata as one unit, so the two teams' releases stay tied together.
- C. Change sets offer no versioning or modular artifacts and scale poorly for two teams sharing one org.
- D. Packages in this model are created and versioned with Salesforce CLI commands, not with point-and-click tools.
Question 28
Universal Containers has decided on a single-org strategy, despite having to deal with the complexity of having multiple lines of business (LOBs) inside a single org.
What are two common challenges in single-org strategy for multiple LOBs? (Choose two.)
- A. Apex design will need to be mature and adhere to strict guidelines to support large enterprise model.
- B. The data model becomes more complex the scope in the org increases.
- C. Making Salesforce work with multiple currencies.
- D. Lack of declarative sharing and visibility capabilities to ensure correct visibility of objects and records.
Show answer and explanation ▾
Correct answer: A, B
When several lines of business share one org, their objects, fields and automation all live together, so the data model grows more complex. Shared objects also mean shared Apex, which needs disciplined design standards so one team's code does not break another's.
Why the other options are wrong:
- C. Multiple currencies are a standard platform feature and are not a challenge specific to a single-org strategy.
- D. Salesforce provides extensive declarative sharing and visibility features, so a lack of them is not the issue.
Question 29
Universal Containers has an active production org and they are planning to release some new features to it next month. The team is working to prepare a deployment plan and reached out to the technical architect for inputs on rollback strategy.
What should a technical architect recommend?
- A. Create a sandbox from production to take the backup of existing metadata. To roll back deployment, use destructivechanges.xml to delete new components and then deploy again to production using metadata from this sandbox.
- B. Backup the existing metadata using ANT Migration Tool. To roll back deployment, manually delete new components and deploy again to production using backed up metadata.
- C. Create a sandbox from production to take the backup of existing metadata. To roll back deployment, manually delete new components and then deploy again to production using metadata from this sandbox.
- D. Backup the existing metadata using the ANT Migration Tool. To roll back deployment, deploy again to production using backed up metadata.
Show answer and explanation ▾
Correct answer: A
A sandbox created from production before the release preserves the previous metadata. Rolling back takes two actions: a destructiveChanges.xml manifest removes the components the release added, and the earlier versions of changed components are redeployed from the sandbox.
Why the other options are wrong:
- B. Deleting new components by hand is slow and error prone compared with a destructive changes manifest.
- C. Deleting new components by hand is slow and error prone compared with a destructive changes manifest.
- D. Redeploying old metadata alone does not remove components that the release added.
Question 30
Cloud Kicks (CK) is launching a new sneaker line during the upcoming holiday season and needs to do a thorough batch data testing before GoLive. CK is using Salesforce unlimited edition. What two sandbox types should the architect recommend for batch data testing? (Choose two.)
- A. Full sandbox
- B. Developer Pro sandbox
- C. Developer sandbox
- D. Partial Copy sandbox
Show answer and explanation ▾
Correct answer: A, D
Batch data testing needs realistic record volumes. A Full sandbox copies all production data and a Partial Copy sandbox copies a sample of production data defined by a sandbox template, so both provide real data to run the batches against.
Why the other options are wrong:
- B. A Developer Pro sandbox copies metadata only and contains no production data.
- C. A Developer sandbox copies metadata only and has the smallest storage allowance.
Question 31
Universal Containers has a highly customized Salesforce org, with many different pieces of configuration and code.
Which configuration item should be covered by executable tests?
- A. Active Process Builders
- B. Validation Rules
- C. Workflow Rules
- D. Case Assignment Rites
Show answer and explanation ▾
Correct answer: A
Active processes built in Process Builder run complex logic on record changes: they can update related records and call Apex or flows. Executable tests that trigger them confirm that logic still works and that it does not hit limits when combined with other automation.
Why the other options are wrong:
- B. Validation rules are simple declarative checks and are not the item singled out for executable tests here.
- C. Workflow rules perform simple declarative actions and are not the item singled out for executable tests here.
- D. Case assignment rules are routing configuration and are not the item singled out for executable tests here.
Question 32
The CTO at Universal Containers decided to implement the Scrum framework for its agile teams, and communicated a set of Serum principles to the company.
Which describes a Scrum principle?
- A. Respect other teams by not doing their work (a developer should not test the software).
- B. Create transparency by being honest and clear about timing, planning, and obstacles.
- C. Embrace change by working on a different scope every day.
- D. Deliver working software, so if a software component is working, avoid changing it.
Show answer and explanation ▾
Correct answer: B
Transparency is one of the pillars of Scrum, together with inspection and adaptation. The team makes its progress, plans and obstacles visible so that everyone can inspect the real state of the work and adjust.
Why the other options are wrong:
- A. Scrum teams are cross-functional and share responsibility for the work, including testing.
- C. Scrum keeps the sprint goal stable during a sprint, so the scope does not change every day.
- D. Scrum encourages continuous improvement, so working software is still changed when that adds value.
Question 33
Universal Containers has discovered a Sev0 defect in production. Tens of thousands of records will be created with incorrect data in minutes, producing significant brand damage as a consequence. The Salesforce administrator has suggested that the defective text field be replaced with a new picklist field directly in production. The page layout will be modified so that the text-field is removed and the new picklist field added.
What should the Salesforce architect advise?
- A. Explain that only developers are certified to make changes directly in production.
- B. Deny the suggestion and explain to everyone that the risk is too high and the next release window is on the weekend.
- C. Call the security team and begin organizing a penetration test.
- D. Pair with the administrator, and review each change as it happens
Show answer and explanation ▾
Correct answer: D
A defect that is corrupting thousands of records within minutes justifies an immediate fix in production. Pairing with the administrator and reviewing each change as it is made keeps a second set of eyes on the work and reduces the risk of the emergency change.
Why the other options are wrong:
- A. No certification rule restricts production changes to developers.
- B. Waiting for the weekend release window lets the incorrect data and brand damage keep growing.
- C. A penetration test addresses security weaknesses and does nothing for a data defect.
Question 34
Which two options should be considered when making production changes in a highly regulated and audited environment? (Choose two.)
- A. All changes including hotfixes should be reviewed against security principles.
- B. After deployment, the development team should test and verity functionality in production.
- C. Any production change should have explicit stakeholder approval.
- D. No manual steps should be carried out.
Show answer and explanation ▾
Correct answer: A, C
In a regulated and audited environment every production change needs a traceable approval from the responsible stakeholders. Every change, including an urgent hotfix, also goes through a security review so that speed never bypasses the controls auditors expect.
Why the other options are wrong:
- B. Verification in production is typically done by business or release staff with proper separation of duties, not by the development team.
- D. Some manual steps are unavoidable, and they are acceptable when documented and approved.
Question 35
Which two actions will contribute to an improvement of code security? (Choose two.)
- A. Use two developers to review and fix current code vulnerabilities.
- B. Implement a pull request and secure code review.
- C. Hire a company specialized in secure code review the current code.
- D. Integrate a static code security analysis tool in the CI/CD process.
Show answer and explanation ▾
Correct answer: B, D
Pull requests with a secure code review put every change in front of a reviewer before it is merged, and a static security analysis tool in the pipeline scans every build automatically. Both make security checks a permanent part of the process.
Why the other options are wrong:
- A. A one-time fix of current vulnerabilities by two developers does not stop new ones from being introduced.
- C. A one-time external review covers only the current code and does not improve the ongoing process.
Question 36
Universal Containers (UC) has multiple development teams that work on separate streams of work, with different timelines. Each stream has different releases of code and config, and the delivery dates differ between them.
What is a suitable branching policy to recommend?
- A. GitHub flow
- B. Trunk-based development
- C. Scratch-org-based development
- D. Leaf-based development
Show answer and explanation ▾
Correct answer: A
In GitHub flow each stream of work lives on its own branch created from the main branch and is merged through a pull request when it is ready. Teams with different delivery dates can therefore release independently of each other.
Why the other options are wrong:
- B. Trunk-based development has every team commit to one shared line, which suits a single shared release cadence.
- C. Scratch-org-based development describes the environment used for development and is not a branching policy.
- D. Leaf-based development is not a recognized branching policy.
Question 37
Universal Containers (UC) is embarking on a large program of work, with different projects and different vendors. UC created a center of excellence (COE) that is struggling with scope creep between the different projects.
What role should the architect suggest be added to the COE?
- A. Release managers
- B. Scrum master
- C. Change managers
- D. Product owner
Show answer and explanation ▾
Correct answer: B
A scrum master keeps teams working to the agreed framework, protects the committed scope of each sprint and removes obstacles between teams. Adding that role to the center of excellence gives it someone accountable for stopping uncontrolled scope growth across projects.
Why the other options are wrong:
- A. Release managers coordinate what is deployed and when, not what each project takes into scope.
- C. Change managers focus on user adoption and communication, not on controlling project scope.
- D. A product owner prioritizes the backlog of one product and does not coordinate scope across separate projects and vendors.
Question 38
Cloud Kicks is switching to Salesforce from a different CRM. They have existing datasets for all standard Salesforce objects.
In which optimized order should the architect recommend these objects be loaded?
- A. Accounts, Contacts, Leads, Products, Opportunities, Opportunity Line Items
- B. Accounts, Contacts, Opportunities, Products, Opportunity Line Items, Leads
- C. Leads, Contacts, Accounts, Opportunities, Products, Opportunity Line Items
- D. Leads, Accounts, Contacts, Products, Opportunities, Opportunity Line Items
Show answer and explanation ▾
Correct answer: B
Parent records are loaded before the records that look up to them. Accounts come first, then contacts and opportunities that reference accounts, then products, then opportunity line items that need both an opportunity and a product. Leads depend on none of these and load last.
Why the other options are wrong:
- A. Loading leads in the middle of the account-related objects is not the recommended sequence.
- C. Contacts cannot be linked to their accounts when they are loaded before accounts.
- D. Starting with leads delays the account hierarchy that most other objects depend on.
Question 39
Universal Containers wants to introduce data volume testing to resolve ongoing performance defects earlier in the lifecycle. Regulations prohibit the use of production data in non-production environments.
Which two options can the architect recommend? (Choose two.)
- A. Request a partial Sandbox copy after the next Salesforce release.
- B. Generate mock data that mimics production data shape and volume.
- C. Use Query Analyzer in production.
- D. Perform data masking on full sandbox after a refresh.
Show answer and explanation ▾
Correct answer: B, D
Volume testing needs data with a realistic shape and size, but regulations keep real production data out of test environments. Generated mock data meets both needs, and so does a Full sandbox whose sensitive fields are masked right after each refresh.
Why the other options are wrong:
- A. A Partial Copy sandbox still contains unmasked production data and only a sample of the volume.
- C. Analyzing queries in production does not provide a non-production environment for volume testing.
Question 40
What are three advantages of using the SFDX? (Choose three.)
- A. Can quickly deploy metadata using Execute Anonymous.
- B. Can create scratch orgs.
- C. Can install application metadata from a central repository.
- D. Can store code on a local machine, or a version control system.
- E. Can use native Deployment Rollback Too to quickly revert to prior state.
Show answer and explanation ▾
Correct answer: B, C, D
Salesforce DX lets developers create disposable scratch orgs, keep source in local project files that sync with a version control system, and install application metadata such as packages into an org from that central source.
Why the other options are wrong:
- A. Execute Anonymous runs Apex code and does not deploy metadata.
- E. Salesforce does not provide a native deployment rollback tool.
Question 41
There are many types of quality assurance techniques that can help minimize defects in software projects.
Which two techniques should an architect recommend, for Universal Containers to incorporate into its overall CI/CD pipeline? (Choose two.)
- A. Automated browser testing
- B. Stress testing
- C. Business verification testing
- D. Static code quality analysis
Show answer and explanation ▾
Correct answer: A, D
Automated browser tests and static code quality analysis both run without human involvement, so a pipeline can execute them on every commit or build and stop a defect before it moves to the next environment.
Why the other options are wrong:
- B. Stress testing is a heavy, occasional exercise and is not run as a routine pipeline step.
- C. Business verification testing is performed manually by business users and is not an automated pipeline step.
Question 42
Universal Containers’s architect is documenting the application lifecycle management (ALM) process to communicate it to the development teams from different implementation partners.
Which three steps apply to any Salesforce development project? (Choose three.)
- A. Test
- B. Develop
- C. Continuous Integration
- D. Build Release
- E. Change Sets
Show answer and explanation ▾
Correct answer: A, B, D
Every Salesforce project follows the same application lifecycle steps regardless of tooling: plan the release, develop, test, build the release, test the release and then release it. Develop, test and build release are three of those steps.
Why the other options are wrong:
- C. Continuous integration is an optional practice that some teams adopt, not a step in every project.
- E. Change sets are one deployment tool among several, not a lifecycle step.
Question 43
A team has completed a sprint and intends to deploy these changes after business approval, but they will immediately begin the next sprint.
What strategy should an architect recommend?
- A. Using Git, create a release branch from the develop branch. All fixes must be made in the release branch. After deployment, merge release with develop.
- B. The first task of the new sprint must be the deployment approval. After that, the other tasks of the sprint can be performed in the environments and Git.
- C. Commit upcoming changes to the features branch without merging into the develop branch. Deploy from the develop branch and then merge new sprit features into the develop branch.
- D. Migrate the current code to the UAT sandbox. Begin new sprint development in the Dev sandbox. Make fixes in the UAT environment and deploy UAT for production after business approval.
Show answer and explanation ▾
Correct answer: A
A release branch created from develop freezes the finished sprint for approval and deployment while new sprint work continues on develop. Fixes found before go-live are made on the release branch, and merging it back afterward keeps develop complete.
Why the other options are wrong:
- B. Making approval a sprint task blocks the team instead of letting the next sprint start immediately.
- C. Holding new work out of develop removes continuous integration of features and complicates later merges.
- D. Fixing directly in the UAT sandbox bypasses version control, so those fixes can be lost.
Question 44
Universal Containers (UC) have developed a managed package targeted for AppExchange. The product includes some Apex code to customize and create layouts. UC is in the testing phase of the package, so it’s not certified yet. During testing on the target org, the Apex code for the layouts fails.
Why are the Apex classes not able to access the metadata of the target org during testing?
- A. Apex Settings to allow the access to metadata is not switched on.
- B. UC needs to get the managed package certified by the Salesforce security review.
- C. The solution is flawed. UC should utilize the Tooling API from a web service call to modify the layouts.
- D. UC needs to turn on Apex Settings within the custom metadata type.
Show answer and explanation ▾
Correct answer: A
Apex in a managed package that has not passed security review can deploy metadata in a subscriber org only when the org's Apex Settings option to deploy metadata from non-certified package versions via Apex is enabled. With it off, the layout code fails.
Why the other options are wrong:
- B. Certification is not required for testing, because the Apex Settings option allows non-certified packages to deploy metadata.
- C. Apex can read and update layouts through the Metadata namespace, so the design is valid.
- D. The option is in Apex Settings in Setup, not inside a custom metadata type.
Question 45
A developer was trying to retrieve the metadata from an org and ran the sfdx force:source:retrieve command.
When the command was run, the developer received the error message: “This command is required to run from within an SFDX project.”
What can be two possible reasons that caused this problem? (Choose two.)
- A. The developer hadn’t run the sfdx force:project:create command.
- B. The developer created the project within VSCode, but ran the command in a separate terminal.
- C. The developer created the project, but ran the command outside of the project directory.
- D. The developer forgot to add the -n option with a project name as a command line argument.
Show answer and explanation ▾
Correct answer: A, C
Source commands look for an sfdx-project.json file in the current directory or one above it. The error appears when no project has been created yet or when the command is run from a folder outside the project.
Why the other options are wrong:
- B. A separate terminal works as long as its working directory is inside the project.
- D. The retrieve command does not take a project name argument, so a missing -n option does not cause this error.
Question 46
Universal Containers (UC) is embarked on an enterprise salesforce transformation journey, UC would like to streamline and automate deployment to different sandboxes during the build phase. Upon customer acceptance in UAT, the company requested to automate the production deployment as well.
As the deployment architect, what is the recommendation to satisfy the customer requirements?
- A. Recommend using SFDX and documents the deployment commands with steps to be executed for each environment.
- B. Recommend using the Continues integration and the Continues deployment tool and build the pipeline to deploy to sandboxes and production.
- C. Recommend using an AppExchange solution that packages the deployment components and you can run the deployment wizard to track deployment result.
- D. Recommend using the ANT script and build a custom application to run the script and use change sets to deploy supported metadata.
Show answer and explanation ▾
Correct answer: B
A continuous integration and continuous deployment tool runs a pipeline that deploys automatically to each sandbox as changes are merged and then, once UAT is accepted, promotes the same validated changes to production. That automates the whole path the customer described.
Why the other options are wrong:
- A. Documented commands that someone runs by hand in each environment are not automation.
- C. A deployment wizard is operated manually for each deployment and does not automate the path.
- D. Change sets cannot be scripted, and a custom application around ANT adds maintenance without a full pipeline.
That was 46 of 118.
The full Salesforce Development Lifecycle and Deployment Architect pack has all 118 questions, each with the answer, the explanation and why the other options are wrong, plus a questions-only copy for timed runs. $39.00 USD, paid once, with free monthly updates and a pass-or-your-money-back guarantee.
Get the full packNot ready yet? Get the first 46 free as one PDF by email. Get the Half Pack →
About the Development Lifecycle and Deployment Architect exam
These 46 Salesforce Development Lifecycle and Deployment Architect questions are free here, and as a free PDF by email.
They are drawn from the same bank as the full Development Lifecycle and Deployment Architect pack, which has 118 questions in total.