SPDLAD

Free SPDLAD practice questions

10 questions from our SPDLAD practice tests, each with the correct answer and a written explanation. Read them with the answers shown, or switch to quiz mode and answer first.

Back to the SPDLAD practice exams

Question 1 of 10, Application Lifecycle Management

A nonprofit with one admin and occasional consultant projects asks whether it should adopt the package development model like a large enterprise it read about. The org has few customizations and no developers on staff. What should the architect advise?

  1. AUse the org development model with source control, which adds traceability without package design overhead Correct answer
  2. BKeep using change sets only, because source control requires a team of at least five developers to be useful
  3. CAdopt second-generation managed packages, because packaging is the only model that Salesforce still supports
  4. DAdopt unlocked packages for every component immediately, because each component then gets its own version number

Answer and explanation: For a small team with limited customization, the org development model with source control gives traceability and repeatable deployments without designing package boundaries and dependencies. Managed packages target ISVs, and change sets remain supported, so packaging is not mandatory. Source control is useful even for one person. Packaging every component individually adds heavy overhead with little benefit for a small org.

Question 2 of 10, Planning

A company has 40 developers, one Full sandbox, two Partial Copy sandboxes, and plenty of Developer sandboxes. Teams fight over the Full sandbox for UAT, performance testing, and training. Which plan BEST uses the available environments?

  1. AUse the Full sandbox for all three activities on a rotating weekly schedule managed by the release manager
  2. BKeep the Full sandbox for staging and performance tests; run UAT and training in the Partial Copy sandboxes Correct answer
  3. CRun UAT and training in Developer sandboxes loaded with sample data and use the Full sandbox only for staging
  4. DRequest a second Full sandbox and dedicate one Full sandbox to UAT and the other to performance testing

Answer and explanation: Staging and performance testing need production-scale data and configuration, which only the Full sandbox provides, while UAT and training work well with representative subsets in Partial Copy sandboxes. Rotating the Full sandbox weekly blocks each activity for weeks and causes conflicts. Developer sandboxes have little storage and no production data, which makes UAT unrealistic. Buying another Full sandbox adds cost when existing environments can cover the need.

Question 3 of 10, System Design. Choose 2.

A financial services company operates three Salesforce orgs in different regions. Each org shares 70% of its configuration but has local regulatory customizations. The company wants to reduce duplicated effort when shipping shared changes. Which approaches should the architect recommend? (Choose two.)

  1. APackage the shared configuration into versioned unlocked packages that are installed in all three orgs Correct answer
  2. BUse change sets from the first region's production org to push shared changes to the other two production orgs
  3. CKeep each region's local customizations in its own package or source directory that depends on the shared package Correct answer
  4. DMerge the three orgs into one org so that shared changes only have to be deployed a single time
  5. EGive each region its own copy of the shared metadata in separate repositories so teams can change it freely

Answer and explanation: Packaging shared configuration once and installing the same version in all orgs removes duplicated delivery work, while regional extensions that depend on the shared package keep local rules isolated. Merging orgs is a large consolidation project that may conflict with regional regulations and is not required to reduce duplication. Change sets only work between orgs connected to the same production org, so they cannot move changes between separate production orgs. Separate copies of shared metadata recreate the duplication the company wants to remove.

Question 4 of 10, System Design

A team wants to enable or disable new automation in production without a deployment, so features can be switched on per business unit after the release. Which design should the architect recommend?

  1. ADeactivate the flows and triggers in production and ask an admin to activate them when the business is ready
  2. BKeep the new automation in a separate sandbox until the business unit asks for it, then deploy it with a change set
  3. CStore the on/off value in a custom label so that translations can switch the feature for each business unit
  4. DCheck a custom metadata type record or custom permission inside the automation before running the new logic Correct answer

Answer and explanation: Custom metadata records and custom permissions act as feature toggles that can be changed without code deployment and can target users or business units. Deactivating triggers in production requires deploying the trigger again, and toggling flows per business unit is not possible. Holding changes back in a sandbox delays deployment and adds merge risk. Custom labels are meant for translatable text, not per-unit configuration, and cannot vary by business unit.

Question 5 of 10, Building

A team using a feature-branch workflow finds that long-lived feature branches cause painful merge conflicts at the end of each sprint. What should the architect recommend?

  1. AKeep feature branches short-lived and merge them into the main integration branch frequently, at least every few days Correct answer
  2. BLock files in the repository so that only one developer at a time can edit a metadata file during the sprint
  3. CStop using pull requests and allow developers to push directly to the release branch to avoid waiting for reviews
  4. DHave each developer work directly in the shared sandbox and retrieve the changes into Git at the end of the sprint

Answer and explanation: Merging small changes frequently keeps branches close to the integration branch, so conflicts stay small and are found early. File locking slows the team and does not fit distributed Git workflows. Working directly in a shared sandbox loses isolation and makes the repository an afterthought. Skipping pull requests removes code review and quality gates, which increases defects.

Question 6 of 10, Building

A team wants every developer to start from an identical, disposable org that has the right features and settings enabled for the project. What should they use?

  1. APartial Copy sandboxes created from a template that includes the main configuration objects
  2. BDeveloper sandboxes cloned each morning from the integration sandbox used by the whole team
  3. CScratch orgs created from a shared scratch org definition file stored in the repository Correct answer
  4. DDeveloper Edition orgs that each developer signs up for and configures by following a checklist

Answer and explanation: A scratch org definition file in source control defines edition, features, and settings so every scratch org starts identically and can be discarded. Cloning sandboxes daily is slow and still drifts. Manually configured Developer Edition orgs are inconsistent and not connected to the company's Dev Hub. Partial Copy sandboxes copy data, refresh only every 5 days, and are a limited resource.

Question 7 of 10, Deploying

An admin team deploys a profile from a sandbox to production with the Metadata API, and afterwards users lose access to several fields that were not part of the release. The deployment manifest included the profile and two new custom objects. What is the MOST likely cause, and how should future releases avoid it?

  1. ADeploying any profile resets field-level security for every object in the org; deploy profiles only through change sets from now on
  2. BThe Metadata API deploys profiles as complete replacements of production profiles; retrieve the full production profile before each deployment
  3. CProfiles cannot be deployed with the Metadata API, so an admin must recreate field permissions manually after each release
  4. DProfile permissions in the source file differed for components included in the manifest; manage access with permission sets deployed alongside the release Correct answer

Answer and explanation: When a profile is deployed, permissions are set for the components present in the deployment, so a stale or partial source profile can change access unexpectedly. Moving access to permission sets keeps releases explicit and small. A profile deployment does not reset every object in the org; it affects the components included. Change sets do not solve the problem since they deploy profile settings for included components in a similar way. Profiles can be deployed with the Metadata API.

Question 8 of 10, Testing

An architect must plan testing for an integration where Salesforce publishes platform events consumed by an external order system. Which test approach provides the BEST coverage?

  1. AApex tests that publish events and assert that the external system has received them by calling it directly from the test
  2. BUnit tests for the external system only, because the publish call in Salesforce always succeeds without errors
  3. CManual testing in production after go-live, since platform events cannot be tested anywhere except production
  4. DApex tests with Test.getEventBus().deliver(), plus end-to-end tests in a sandbox wired to the order system's test instance Correct answer

Answer and explanation: Apex tests can deliver test events to subscribers in the org to verify publishing and subscribing logic, and end-to-end tests with the external test system validate the integration. Apex tests cannot make real callouts. Platform events work in sandboxes, so production-only testing is unnecessary and risky. Publishing can fail, for example on limits or validation, so Salesforce-side tests are needed.

Question 9 of 10, Releasing. Choose 2.

A company has several Salesforce orgs that must receive the same managed configuration package each release. Releases to some orgs fail because those orgs have local changes conflicting with the package. Which TWO practices help? (Choose two.)

  1. ADefine clear boundaries so local customizations extend, rather than change, packaged components Correct answer
  2. BInstall and test each release in a sandbox of every target org before production Correct answer
  3. CAllow each org's admins to edit packaged components directly when they need changes
  4. DSkip releases for any org that has conflicts until its admins remove all local changes
  5. EDeploy the package with change sets to bypass the conflict checks in each target org

Answer and explanation: Keeping local work as extensions of packaged components avoids conflicts, and testing each release in each org's sandbox catches remaining issues before production. Editing packaged components locally causes the conflicts. Skipping releases leaves orgs behind. Change sets cannot move metadata between separate production orgs and do not bypass conflicts.

Question 10 of 10, Operating

Production support sees occasional Apex governor limit exceptions in a batch job that processes nightly loads. The volume of incoming data is growing 20 percent per quarter. What should the operations plan include?

  1. ARestart the batch job manually each time it fails and accept occasional failures as a normal cost
  2. BMonitor job failures and data growth, and plan capacity reviews that refactor the job before limits are reached Correct answer
  3. CIncrease the batch size to the maximum allowed value so the job finishes with fewer transactions each night
  4. DAsk Salesforce Support to raise the governor limits for the org because the data volume keeps growing

Answer and explanation: Proactively monitoring failures and growth trends allows the team to optimize the job before volume makes failures routine. Larger batch sizes increase per-transaction work and make limit errors more likely. Governor limits are fixed per transaction and are not raised on request. Manual restarts accept recurring failures and data delays.