Development-Lifecycle-and-Deployment-Architect Salesforce Certified Development Lifecycle and Deployment Architect (WI25) Question and Answers

Active Process Builders are configuration items that should be covered by executable tests, as they can have complex logic and actions that affect the behavior and performance of the application. Validation Rules, Workflow Rules, and Case Assignment Rules are also configuration items, but they are simpler and less likely to cause errors or conflicts. Therefore, they can be verified by manual testing or declarative tools.

The best practice for writing test classes is to not use SeeAllData and generate data in the test methods. This ensures that the test classes are independent of the data in the org and can run successfully in any environment. Creating test data in production or setting SeeAllData to True can cause unexpected failures or data conflicts.

 Generating a code review checklist to ensure consistency between reviews and different reviewers and conducting a review that combines human efforts and automatic checks by the tool to detect all flaws are two recommended practices to follow when conducting secure code reviews, as they can help improve the quality and security of the code and reduce the risk of vulnerabilities or errors. Focusing on the aggregated reviews to save time and effort, to remove the need to continuously monitor each meaningful change is not a good practice, as it can miss important details or changes that can affect the code functionality or security. Using the code review software as the tool to flag which developer has committed the errors, so the developer can improve is not a constructive way of conducting code reviews, as it can create a blame culture and discourage collaboration and learning. See Secure Coding Guidelines for more details.

The best approaches for UC to consider are B and C, as they both involve agile methodologies that can accommodate changing requirements and deliver value to the end users in an iterative and collaborative way. Option B is a good approach, as it allows UC to scope and estimate the project, identify the minimum viable product, and co-create the design with the project stakeholders using a product backlog and sprints2. Option C is also a good approach, as it allows UC to gather the remaining requirements and complete the architecture and design, then initiate the build/test processes with frequent reviews by the stakeholders, and perform acceptance testing and validate compliance with regulatory requirements3. Option A is not a good approach, as it involves a waterfall project that is rigid and sequential, and may result in re-work and delays if the requirements change or conflict4. Option D is not a realistic approach, as it assumes that budget and timeline will not be a factor with an agile methodology, which is not true, as agile projects still need to have clear scope, budget, and timeline constraints

The best option is to send a package.xml file with associated metadata in a .zip file to the Release Management team, as this will allow them to use the Metadata API to deploy the components to the production org. Option A is not feasible, as change sets cannot be downloaded from sandboxes. Option B is not sufficient, as the readMetadata call only retrieves the metadata information, not the actual components. Option C is not reliable, as the audit trail may not capture all the changes made in the sandbox.

 Design standards can help ensure consistency and quality of the code and configuration across multiple project teams. They can also prevent design conflicts and reduce technical debt.

The reason why the Apex classes are not able to access the metadata of the target org during testing is that UC needs to get the managed package certified by the Salesforce security review. This is because Apex code in a managed package can only access the metadata of the target org if the package has passed the security review and has been granted the Modify Metadata permission. See Apex Metadata API for more details.

 The maximum size of deployed .zip file is 39MB for Metadata API. You can deploy up to 10,000 files, but retrieve more than 10,000 files using Metadata API.

The best options to enforce code quality in UC’s continuous integration process are to introduce manual code review before deployment to the testing sandbox and to introduce static code analysis before deployment to the testing sandbox. Manual code review can help identify and fix any errors, bugs, or best practices violations in the code. Static code analysis can help check the code quality, complexity, and security using automated tools and standards. Introducing manual code review before deployment to the production org may be too late, as the code may have already caused defects or issues in the testing sandbox. Increasing the size of the testing team assigned to the project may not improve the code quality, as the testing team may not have the skills or authority to review or modify the code. Testing data creation is outside the scope of code quality.

 This is the correct answer because the executive sponsor is the senior leader who provides strategic direction, funding, and support for the project. The executive sponsor is also responsible for approving any changes to the project scope, budget, or timeline that may affect the project outcomes or benefits.

An unmanaged package is a suitable option for deploying an application from a contractor’s own Salesforce to UC’s Salesforce environments. An unmanaged package allows the contractor to share the application code with UC, while giving UC full control and ownership of the code. An unmanaged package also does not require a namespace prefix or a security review.

These are the correct answers because they are essential components of a governance framework that helps to manage the changes and enhancements to the Salesforce application. Requirements management is the process of capturing, analyzing, prioritizing, and tracking the business needs and requests. Change control log is the tool that records and tracks the status of all change requests and approvals. Documentation repository is the central place that stores and organizes all the project artifacts and deliverables. Automated testing and continuous integration are not necessary components of a governance framework, but rather best practices for ensuring the quality and reliability of the code and configuration.

Implementing 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 is one of the options an architect can recommend to accomplish the customer 360-degree view. This way, the architect can ensure that there is a single source of truth for the customer data, and that the data is consistent and synchronized across the different orgs. Implementing 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 is another option an architect can recommend, as it allows the architect to create a reusable and scalable solution that can provide a unified view of the customer data from multiple sources. Implementing a Complete Graph multi-org strategy or migrating from multi-org to single-org strategy are not feasible or optimal options, as they can introduce complexity, cost, or risk to the architecture.

 Scratch orgs are the best development environments for the package development model, as they allow developers to start a new project, start a new feature branch, or start automated testing in a disposable and configurable org. Scratch orgs also track all the changes automatically and can be easily integrated with version control systems. Scratch orgs can coexist with other models, such as the org development model, but this is not the main reason to start using them. Dev sandboxes are not suitable for the package development model, as they are not within the code deployment path and they do not support source-driven development.

Three areas that should be tested as part of data migration are: data transformation against source system, contact association with correct account, and account and lead ownership. These areas can help ensure that the data is migrated correctly and accurately, and that the relationships and ownership are preserved. Lead assignment is not an area that should be tested as part of data migration, as it is a business process that is triggered after the data is migrated. Page layout assignments is also not an area that should be tested as part of data migration, as it is a configuration setting that is independent of the data. See Data Migration for more details.

 The Metadata API can be used to migrate configuration changes between two organizations or export current customization in the organization as an xml file. The Metadata API allows developers to retrieve, deploy, create, update, or delete the metadata components that define the structure and behavior of the Salesforce org. The Metadata API cannot perform CRUD operations to manage records in the organization or migrate data changes between two organizations using a csv file. These operations require the use of other APIs, such as the REST API or the Bulk API.

The benefits of managing change with Packaged Development are:

Versioning to help with change management, as it allows the developers to track the changes made to the metadata and roll back to previous versions if needed.

Making the release cycle more efficient and agile, as it enables the developers to deploy smaller and more frequent updates to the orgs, reducing the risk of errors and conflicts.

Modular development process with specification of dependencies among packages, as it allows the developers to break down the metadata into logical units that can be reused and updated independently, while ensuring that the dependencies are resolved correctly. Managing the number of sandboxes needed to successfully deploy is not a benefit of Packaged Development, as it is more related to the sandbox strategy and the development model. Clearly dividing developers and testers is also not a benefit of Packaged Development, as it is more related to the team structure and the testing strategy.

To reduce deployment time, UC can use the following strategies:

Use recent deployment validations and the quick deploy feature, which can allow them to skip the validation step in production and deploy the changes faster, if they have already validated the changes in a sandbox within the last four days.

Specify the test to run by using RunSpecifiedTests test level, which can allow them to run only the tests that are relevant to the components being deployed, instead of running all the tests in the org.

Develop, Build Release, and Test are three steps that apply to any Salesforce development project, regardless of the size, complexity, or methodology. Develop is the step where the developers create the code and configuration components that meet the requirements and specifications. Build Release is the step where the components are packaged and deployed to different environments for testing and validation. Test is the step where the components are verified and validated against the requirements and specifications, as well as the quality and performance standards. Continuous Integration and Change Sets are not steps, but tools or techniques that can be used to facilitate the development project. Continuous Integration is a technique that involves merging the code from different developers frequently and automatically, to ensure consistency and avoid conflicts. Change Sets are a tool that allows the developers to move the components from one org to another org.

 Creating design standards to require help text on all custom fields and custom objects is an option that an Architect should recommend, as it helps to provide context and guidance for the administrators and users on how the fields and objects are used and what they mean. Creating design standards to consistently use the description field on custom fields is also an option that an Architect should recommend, as it helps to document the purpose and function of the fields and make them easier to maintain and understand. Creating design standards to consistently use the description field on custom objects is not an option that an Architect should recommend, as it is not a mandatory field and may not provide enough information for the administrators. Creating design standards with a document to store all custom objects and custom fields is not an option that an Architect should recommend, as it may be difficult to keep the document updated and synchronized with the actual metadata in the org. Creating design standards to require all custom fields on all custom object page layouts is not an option that an Architect should recommend, as it may clutter the user interface and reduce the usability and performance of the application.

A virtual Kanban board tool is a good way to increase transparency and collaboration among agile development teams, as it allows them to visualize their work, track their progress, and identify any bottlenecks or issues. A daily stand-up meeting, a daily email, or a shared spreadsheet are not as effective or efficient as a Kanban board tool.

The only condition that would make continuous integration unsuitable for UC is if their Salesforce instance has only standard functionality, as this would mean that there is no need for code deployment or version control. Option A is not a valid reason, as test scripts can be generated and executed as part of the continuous integration process. Option B is not a valid reason, as a full sandbox is not a requirement for continuous integration, and other types of sandboxes can be used instead. Option D is not a valid reason, as there are free or low-cost software tools available for continuous integration, such as Jenkins, Git, and Ant

The best practice is to fix the issues in the development sandbox, migrate them to the testing sandbox, and deploy to production after testing. This ensures that the development and testing environments are in sync and that the changes are properly tested before going to production. Fixing the issues in the testing environment and deploying them to production may introduce errors or conflicts, as the development environment may not have the same changes. Fixing the issues in the test environment and migrating them to the development sandbox may also cause errors or conflicts, as the development environment may have other changes that are not ready for testing.

 In the post refresh plan, modifying external ids to a known valid set of values for UAT is the best way to resolve the issue of objects that have external ids from production no longer pointing to valid external ids in the UAT environment. This way, the data integrity and consistency across the integrated systems can be maintained. Letting UAT point to production integrations and rolling back each transaction after it finishes is not a good practice, as it can cause data loss or corruption in production. Deleting all the data and using an automated testing tool to create new data across all the systems in UAT is not feasible, as it can take a lot of time and resources. Masking the external id so nobody can see the production value is not a solution, as it does not address the underlying problem of invalid references.

These are the correct answers because the project steering committee plays a role in approving deviations from the chosen methodology, when required to address project issues, and setting the criteria for selecting agile or waterfall methodology to be used on internal projects. The steering committee does not design a methodology, but rather chooses one from the existing ones. The steering committee does not enforce that corporate project stage gates are part of the chosen methodology but rather ensures that the chosen methodology aligns with the corporate project stage gates.

Going to a single-org strategy can provide several benefits, such as lower administration overhead costs, improved Chatter collaboration across different business units, and easier to get a 360-view of the customer. Consolidating the business processes and unifying the data model among all lines of business are not benefits of a single-org strategy, but rather challenges that need to be addressed during the org merge process.

Change sets provide an easy way to deploy related components, as they allow the user to select the components from a list and add them to the change set. They also provide a simple and declarative method for deployment, as they do not require any coding or scripting. Change sets do not provide the ability to deploy a very large number of components easily, as they have a limit of 10,000 components per change set. They also do not provide the ability to track changes to components, as they do not have any version control or history features.

 B and D are the decisions that should be made by an Architecture Review Board (ARB), as they involve choosing the best methodology and authentication mechanism for the project, which have significant impacts on the project scope, quality, and security. A is not a decision that should be made by an ARB, as it is a low-level design decision that can be made by the developers or technical leads. C is not a decision that should be made by an ARB, as it is a testing decision that can be made by the testers or quality assurance leads.

QUESTIONNO: 87

What is the process used to initiate a connection for change sets?

A. Modify the source org to allow an outbound connection to the target org

B. Modify the target org to accept an outbound connection from the source org

C. Modify the target org to accept an inbound connection from the source org

D. Modify the source org to allow an inbound connection to the target org

Answer: A

A is the correct answer, as the process to initiate a connection for change sets is to modify the source org to allow an outbound connection to the target org. B and C are incorrect, as the target org does not need to accept any connection from the source org, but only to deploy the inbound change sets that are sent from the source org. D is incorrect, as the source org does not need to allow any inbound connection to the target org, but only to create and upload the outbound change sets to the target org. You can learn more about this process in the Change Sets Basics module on Trailhead.

 To address the bug in production, the architect should recommend the following approaches: Investigate potential data impacts, fix the bug in a hotfix branch, and seek stakeholder approval for the hotfix. Investigating the data impacts can help assess the severity and scope of the bug, and determine the best course of action to mitigate any data loss or corruption. Fixing the bug in a hotfix branch can help isolate the change and avoid introducing any other changes from the current sprint. Seeking stakeholder approval for the hotfix can help ensure compliance with the regulatory requirements and governance policies.

The product team should review the Checkmarx errors and determine if they need to fix them or not. If the errors are false positives, meaning that they do not indicate a real security issue, the product team should mark them as such and attach an explanation, then submit the security review. This will help the Salesforce security review team to understand the rationale behind the code and avoid unnecessary rejections. The product team should not leave the errors to the Salesforce security review team, as they may reject the request if they find any potential security issue. The product team should not leave a partner support case, as this is not the proper channel for resolving code issues. The product team should not fix all the errors before submitting the security review, as some of them may not be relevant or critical, and fixing them may introduce new bugs or delays.

 The two actions that can effectively contribute to the quality and agility of the team are: Define and follow code standards, and use static code analysis tool in the pipeline before manual code review. Code standards can help ensure consistency, readability, and maintainability of the code, as well as reduce errors and bugs. A static code analysis tool can help automate the code review process and identify any issues or violations of the code standards before the manual review. Choosing the most senior developer to help the technical lead or skipping the code review are not effective actions, as they can lead to more errors and delays.

 Three advantages of using the SFDX are: can store code on a local machine, or a version control system; can create scratch orgs; and can install application metadata from a central repository. These advantages can help improve the development experience, as they allow developers to work with source-driven development, use ephemeral and configurable environments, and access metadata components from a shared location. Can quickly deploy metadata using Execute Anonymous is not an advantage of using the SFDX, as Execute Anonymous is a feature of the Developer Console that allows running Apex code, not deploying metadata. Can use native Deployment Rollback Tool to quickly revert to prior state is also not an advantage of using the SFDX, as there is no such tool in the SFDX. See [Salesforce DX] for more details.

 To reduce defects, the architect should recommend increasing the code coverage to 95% and adding more details to the acceptance criteria. Increasing the code coverage can help ensure that the code is thoroughly tested and meets the quality standards. Adding more details to the acceptance criteria can help clarify the requirements and expectations for the functionality. Using the test script to define the test classes or having the test analyst create the test classes are not good practices, as they can introduce errors and inconsistencies in the testing process.

The best option is to ensure the sandbox will be on the release preview, which means it will be upgraded to the new platform release before the production environment. This will allow the team to test the deployment in a realistic scenario and catch any potential issues. Option A is not ideal, as the staging and production support sandboxes will not be on the same platform version as the production environment. Option C is also not ideal, as the production support sandbox will not be on the same platform version as the production environment. Option D is risky, as the staging sandbox refresh may not complete in time for the go-live.

According to the [Salesforce Performance Testing Guide], performance tests must be run in a sandbox, a business justification must be provided to Salesforce, and performance tests may be run without advanced notice, but Salesforce will not store performance logs. Salesforce does not need to be informed 7 days before starting performance tests, and Salesforce will not monitor test activity to ensure there are no issues with Salesforce Services.

This is the correct answer because Salesforce prohibits stress testing against production because it is a shared environment that hosts multiple customers and tenants. Stress testing against production may affect the performance and availability of the Salesforce service for other customers and violate the terms of service. Stress testing should be done in a sandbox or a developer edition org that is isolated from production.

Minor and Major releases should use the same Full copy sandbox for UAT, as this will ensure that the testing environment is consistent and has the same data and configuration as the production org.

Maintaining separate developer branches for minor and major releases, maintaining a single repository for applications with individual branches for projects, and maintaining a single-entry point for production from the master branch are three best practices that UC should consider for version control. These practices help to:

Isolate and manage different types of changes in separate branches, such as bug fixes, enhancements, or new features1.

Organize and track the code and configuration for each application and project in a logical and consistent way2.

Ensure that the master branch reflects the current state of production and that only tested and approved changes are deployed to production

 D is the correct answer, as the Agile methodology is the best way to handle a project with a fixed deadline and a changing scope, such as retiring a business-critical application that will no longer be supported. The Agile methodology allows for frequent feedback, iterative development, and adaptive planning, which can help to deliver a working solution on time and within budget. A is incorrect, as executive leadership is not a methodology, but a role that can support the project. B is incorrect, as a requirements traceability matrix is a tool that can help to track the requirements and their fulfillment, but not a methodology. C is incorrect, as a business continuity plan is a document that outlines how to resume business operations in the event of a disruption, but not a methodology. You can learn more about the Agile methodology in the Agile Basics module on Trailhead.

A change to a multi-org strategy can be justified by two reasons: launching a new line of business with independent processes and acquiring a company that has its own Salesforce org and operates in a different business with its own set of regulatory requirements. These reasons indicate that the single-org strategy is no longer feasible or optimal, as it would require too much customization, complexity, and compliance. Using Chatter for collaboration among different business units is not a reason to change to a multi-org strategy, as Chatter can work across multiple orgs. Following a unification enterprise architecture operating model is also not a reason to change to a multi-org strategy, as this model implies having orgs with the same processes implemented for each business unit, which is more suitable for a single-org strategy.

Testing load considering 50% more call volume is a reasonable test load consideration, as it can simulate the expected future growth and ensure that the application can handle the increased demand without compromising the performance or functionality.

A single-org strategy is the best option for UC, as it will enable them to achieve business process standardization among all business units, which is the main goal of the CEO. A multi-org strategy would make it harder to enforce consistent processes and policies across the different business units, and would also increase the costs and complexity of managing multiple orgs. A common managed package could help with some aspects of standardization, but it would not cover all the possible scenarios and customizations that UC might need.

The recommended strategy for deploying a large volume of metadata is to identify metadata dependencies, create logical groupings, and deploy in the right order. This can help avoid errors and ensure a smooth deployment process.

 The Metadata API is mainly used for file-based deployment, such as deploying an application from a Developer Edition org to the AppExchange, or from a sandbox org to a production org. The Ant Migration Tool is a wrapper around the Metadata API, so it is not a direct use case for writing code to call retrieve() and deploy() methods. The Salesforce CLI (SFDX) uses the Source-Driven Development model, which relies on the source code as the source of truth, rather than the Metadata API.

The architect should suggest the following actions to help UC discover the failing unit tests earlier: Train the administrators to make their changes in a special “admin changes” sandbox, and then promote to production, and ensure a metadata backup is committed to version control every day and a diff published to the release team. Training the administrators to use a sandbox can help prevent direct changes to production that can break the code or the unit tests. Ensuring a metadata backup and a diff can help identify any changes that have been made to production and compare them with the source of truth.

Waterfall development methodology has some positive aspects, such as:

Complex processes that will need to be built are thoroughly understood and documented before coding begins, which can reduce the risk of ambiguity and rework.

Milestones, timelines and estimates tend to be more accurate and predictable due to the upfront due diligence, which can help with project management and stakeholder expectations.

 An application with many systems and inter-dependencies between components is a project situation that favors a waterfall methodology, as it requires a high level of planning and design upfront to ensure the integration and compatibility of the components. An application with regulatory compliance requirements to be validated by outside agencies is also a project situation that favors a waterfall methodology, as it requires a clear and detailed documentation of the requirements and specifications, as well as a formal and rigorous testing and validation process. An application in post-production, with incremental changes made by a small team is a project situation that favors an agile methodology, as it allows for faster and more frequent delivery of changes, as well as more flexibility and collaboration. An in-house application with a fixed team size, but an open timeline and flexible requirements is also a project situation that favors an agile methodology, as it allows for more creativity and experimentation, as well as more feedback and adaptation.

A workflow and a trigger are considered minor changes that do not require alignment with a Salesforce major or patch release. However, they still need proper testing, training, and change management before being deployed to production. Therefore, the architect should plan this as a minor release with training and change management.

 The test class will fail and the developer will see a message stating: Methods defined as TestMethod do not support Web service callouts. This is because the test class cannot make a real web service callout, as it would depend on an external service that may not be available or reliable. Therefore, the developer needs to use the Test.setMock() method to specify a mock class that simulates the web service response. The test class will not skip the web service callout, nor will it make the actual callout. It is possible to save the test class without the Test.setMock() statement, but it will not run successfully.

C and D are the best approaches to develop customizations once and make them available in both orgs, as they use packages that can be installed and updated in multiple orgs. A is not a good approach, as it does not allow for versioning and dependency management of the customizations. B is not a good approach, as it does not support deploying metadata components between orgs.

 A single org strategy has the benefits of increased ability to collaborate, centralized data location, and consistent processes across the business. These factors can improve efficiency, data quality, and user adoption. A single org strategy may not allow for tailored implementation or fewer inter-dependencies, as different business units may have different requirements and dependencies.

B and D are the correct answers, as partner developer edition orgs and developer edition orgs are the appropriate environments for creating a managed package. A managed package is a collection of application components that are locked and versioned, and can be installed and upgraded in multiple orgs. A is incorrect, as developer pro sandbox orgs are not suitable for creating a managed package, as they are copies of a production org that are used for development and testing purposes. C is incorrect, as production orgs with LMA are not suitable for creating a managed package, as they are used for managing the licenses and subscribers of a managed package. You can learn more about this topic in the Package Development Model module on Trailhead.

 Developing a new Customer Community with an Agile methodology and a three-month timeline poses some risks, such as the functionality that can be delivered in 3 months is unknown, and may not meet the needs of the business, and given the lack of requirements and the three-month timeline commitment, the project may not be feasible. Agile is an appropriate development methodology for Customer Community implementations, as it allows for iterative and incremental delivery of value. Agile also allows for changes to requirements, so customers can provide feedback and prioritize the most important features.

The usual symptoms of inadequate negative tests are the ones that indicate that the code is not handling the possible errors or exceptions that may occur in different scenarios. For example, when an Apex batch job runs at a scheduled time, an increased number of Apex execution errors occur over all1, this means that the code is not handling the possible concurrency issues, governor limits, or data quality issues that may arise when the job runs. Similarly, when an Apex process runs into an un-handled exception when an HTTP callout has an unexpected status code in the response body1, this means that the code is not handling the possible network issues, authentication issues, or API errors that may occur when making the callout. Another symptom is when developers constantly ask the testers for a screenshot of the error and the exact steps of reproducing the error1, this means that the code is not logging the error details or providing meaningful error messages to the users. Option A is incorrect because turning to the debug log for details of the failed Apex executions is not a symptom of inadequate negative tests, but a normal practice of debugging the code. Option E is incorrect because the delivered user interfaces are not meeting the expectations of the business users is not a symptom of inadequate negative tests, but a symptom of poor user interface design or incomplete requirements.

 To support application lifecycle methodology, you need tools that can help you manage the source code, automate the deployment process, and track the issues and bugs. A version control repository is a tool that allows you to store, track, and collaborate on the source code of your application. A continuous integration tool is a tool that allows you to automate the deployment of your code to different environments, as well as run tests and validations. An issue tracking tool is a tool that allows you to record, monitor, and resolve the issues and bugs that arise during the development and testing phases. A database management system is a tool that allows you to store, manipulate, and query data, but it is not directly related to application lifecycle methodology. A middleware is a software layer that facilitates communication and data exchange between different applications, but it is not directly related to application lifecycle methodology either.

 The following issues could be affecting their deployment times:

The number and complexity of Apex tests will have a large impact on the deployment time, as the tests will take longer to run and consume more resources.

Some components such as profiles, custom junction objects, and fields take longer to process than others, as they have more dependencies and validations to check.

Users are working in the org during deployment can affect users and the deployment, as they can lock records or metadata components that are being deployed, causing errors or delays.

Two advantages of automated test data loads over manual data loads are that automated loads can be done with no human oversight and that automated loads are reliable in their results. Automated loads can be scheduled and executed without human intervention, which saves time and effort. Automated loads can also ensure that the data is consistent and accurate, which reduces errors and risks. Automated loads do not necessarily increase costs, as they can also save money in the long run by improving efficiency and quality. Automated loads can also be scripted by CICD tools, which is one of the benefits of using them.

Two software development methodologies that an architect should suggest to solve the issue of visibility are Scrum and Kanban. Scrum is based on the value of openness, which means that the team members and stakeholders share information and feedback regularly and transparently. Kanban is based on the principle of making everything visible, which means that the team uses a visual board to track the progress and status of the work items. Waterfall is not a good methodology for visibility, as it does not allow for frequent communication and feedback. DevOps is not a methodology, but a culture and practice that aims to improve collaboration and delivery across the software development lifecycle.

The best strategy to delete the day’s test data in a partial copy sandbox every night is to execute a batch job that deletes all records created on the day. This way, you can automate the process of cleaning up the test data and avoid manual work. Manually deleting all records individually is time-consuming and error-prone. Creating a new developer copy sandbox every night is unnecessary and wasteful, as developer copy sandboxes have limited storage and functionality. Refreshing the sandbox every night is also not feasible, as partial copy sandboxes have a refresh interval of 5 days.

The development model that is optimized for this need is to use unlocked packages to deploy the global code and allow each country to create its customized unlocked package extensions. This model can help maintain the global standardization and consistency of the code, while also providing the flexibility and autonomy for each region to customize some aspects of the code for their specific needs. Unlocked packages are a type of second-generation packaging that can be created and installed using the Salesforce CLI or APIs, and can be upgraded or uninstalled as needed. Unlocked package extensions are a type of unlocked packages that can depend on and extend another unlocked package or a managed package. See Unlocked Packages for more details.

Managed package is the recommended mechanism for releasing an application that can be installed in other Salesforce environments. Managed packages allow for versioning, licensing, and upgradeability of the application. They also protect the intellectual property of the developer by hiding the source code of the Apex classes and Visualforce pages.

 DOM changes introduced by a release can cause automated browser tests to need updating, as they can affect the way the browser interacts with the web page elements and the selectors used to identify them. New standard fields, metadata schema changes, and new Apex methods are not likely to affect automated browser tests, as they are mostly related to the backend functionality and data model of Salesforce. See [Automated Browser Testing] for more details.

Two testing types that an architect should recommend to ensure users don’t face platform slowdowns during peak business hours are: stress testing and performance testing. Stress testing is the process of testing the system under extreme load conditions, such as a large number of concurrent requests, to determine its breaking point and identify any potential bottlenecks or failures. Performance testing is the process of testing the system under normal or expected load conditions, such as the average number of requests per day, to measure its response time, throughput, and resource utilization. These testing types can help ensure that the web service can handle the expected volume and frequency of requests, and that it meets the performance and reliability requirements. Unit testing is not a testing type that an architect should recommend for this scenario, as it is the process of testing individual units of code, such as methods or classes, to verify their functionality and logic. Load testing is also not a testing type that an architect should recommend for this scenario, as it is the process of testing the system under a specific load condition, such as a fixed number of requests, to evaluate its behavior and performance. See Testing RESTful Web Services for more details.

Announcing the maintenance window ahead of time, defining a rollback strategy, and temporarily suspending configuration changes in production are all best practices for production deployment. These practices help to minimize the risk of errors, downtime, and conflicts during the deployment process.