Exam Bundle

The elastic job runs against the intended target databases and excludes an unapproved database from execution. That result is evidence for another corrective change: Create supported Azure SQL elastic jobs with a job agent, credentials, target groups and the minimum required permissions. The alternative problem is: A company wants the same integrity check applied to hundreds of Azure SQL Database single databases but assumes each hosts SQL Server Agent. The current Oakridge Public Services issue is instead: A database task's schedule is stored as a local engineer's calendar reminder, so it stops running when that employee leaves. Its decisive acceptance evidence would be: The task runs on schedule after the engineer's account is disabled and the operations team receives completion evidence.

An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages. That result is evidence for another corrective change: Record structured task identity, target resource, input version, start and end times and step results in protected logs. The alternative problem is: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. The current Oakridge Public Services issue is instead: A database task's schedule is stored as a local engineer's calendar reminder, so it stops running when that employee leaves. Its decisive acceptance evidence would be: The task runs on schedule after the engineer's account is disabled and the operations team receives completion evidence.

The runbook completes the authorized scale action while source code and job output contain no reusable password. That result is evidence for another corrective change: Use an appropriately scoped managed identity or secure automation credential design and eliminate embedded administrator passwords. The alternative problem is: An Azure Automation runbook performs a database scale operation with a stored SQL administrator password despite supporting managed identity authentication. The current Oakridge Public Services issue is instead: A database task's schedule is stored as a local engineer's calendar reminder, so it stops running when that employee leaves. Its decisive acceptance evidence would be: The task runs on schedule after the engineer's account is disabled and the operations team receives completion evidence.

Configure monitored team-owned alert channels and escalation for task successes, failures and skipped schedules. That relates to a separate SQL administration problem: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. The Oakridge Public Services requirement is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. That approach does not directly establish the required fix; the requested acceptance evidence is: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

Make the task idempotent or record a durable execution checkpoint before allowing retries in the governed workflow. That relates to a separate SQL administration problem: A maintenance automation runbook retries a failed data patch and applies the same update twice because it has no idempotency check. The Oakridge Public Services requirement is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. That approach does not directly establish the required fix; the requested acceptance evidence is: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

Use an appropriately scoped managed identity or secure automation credential design and eliminate embedded administrator passwords. That relates to a separate SQL administration problem: An Azure Automation runbook performs a database scale operation with a stored SQL administrator password despite supporting managed identity authentication. The Oakridge Public Services requirement is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. That approach does not directly establish the required fix; the requested acceptance evidence is: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

The task runs on schedule after the engineer's account is disabled and the operations team receives completion evidence. That result is evidence for another corrective change: Move the work to a managed scheduled automation mechanism with a service identity, ownership metadata and alerting. The alternative problem is: A database task's schedule is stored as a local engineer's calendar reminder, so it stops running when that employee leaves. The current Oakridge Public Services issue is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. Its decisive acceptance evidence would be: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

The job history lists each approved target exactly once and contains no unexpected production database. That result is evidence for another corrective change: Maintain elastic job target groups through a controlled inventory and verify membership and execution scope before each rollout. The alternative problem is: A database elastic job runs on only one test database because the production target group was never updated after adding new databases. The current Oakridge Public Services issue is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. Its decisive acceptance evidence would be: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

The task executes the intended procedure but cannot modify a restricted finance table on a target database. That result is evidence for another corrective change: Create or map a narrowly privileged job identity and grant only required procedure permissions on each approved target. The alternative problem is: A scheduled Azure SQL task retains broad database owner permissions even though it only needs to invoke one maintenance procedure. The current Oakridge Public Services issue is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. Its decisive acceptance evidence would be: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

Create or map a narrowly privileged job identity and grant only required procedure permissions on each approved target. That relates to a separate SQL administration problem: A scheduled Azure SQL task retains broad database owner permissions even though it only needs to invoke one maintenance procedure. The Timberline Foods requirement is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. That approach does not directly establish the required fix; the requested acceptance evidence is: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

Alert on per-target job execution outcomes and record failed database targets rather than using agent availability as proof of success. That relates to a separate SQL administration problem: An elastic job fails on one shard but the automation dashboard reports green because it checks only the job-agent heartbeat. The Timberline Foods requirement is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. That approach does not directly establish the required fix; the requested acceptance evidence is: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

Record structured task identity, target resource, input version, start and end times and step results in protected logs. That relates to a separate SQL administration problem: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. The Timberline Foods requirement is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. That approach does not directly establish the required fix; the requested acceptance evidence is: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

The job history lists each approved target exactly once and contains no unexpected production database. That result is evidence for another corrective change: Maintain elastic job target groups through a controlled inventory and verify membership and execution scope before each rollout. The alternative problem is: A database elastic job runs on only one test database because the production target group was never updated after adding new databases. The current Timberline Foods issue is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. Its decisive acceptance evidence would be: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

The task executes the intended procedure but cannot modify a restricted finance table on a target database. That result is evidence for another corrective change: Create or map a narrowly privileged job identity and grant only required procedure permissions on each approved target. The alternative problem is: A scheduled Azure SQL task retains broad database owner permissions even though it only needs to invoke one maintenance procedure. The current Timberline Foods issue is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. Its decisive acceptance evidence would be: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages. That result is evidence for another corrective change: Record structured task identity, target resource, input version, start and end times and step results in protected logs. The alternative problem is: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. The current Timberline Foods issue is instead: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. Its decisive acceptance evidence would be: A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates.

Configure monitored team-owned alert channels and escalation for task successes, failures and skipped schedules. That relates to a separate SQL administration problem: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. The Timberline Foods requirement is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. That approach does not directly establish the required fix; the requested acceptance evidence is: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

Create or map a narrowly privileged job identity and grant only required procedure permissions on each approved target. That relates to a separate SQL administration problem: A scheduled Azure SQL task retains broad database owner permissions even though it only needs to invoke one maintenance procedure. The Timberline Foods requirement is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. That approach does not directly establish the required fix; the requested acceptance evidence is: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

Add approved dry-run checks, a backup or restore checkpoint and stop-on-failure conditions before the database task proceeds. That relates to a separate SQL administration problem: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. The Timberline Foods requirement is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. That approach does not directly establish the required fix; the requested acceptance evidence is: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

The task runs on schedule after the engineer's account is disabled and the operations team receives completion evidence. That result is evidence for another corrective change: Move the work to a managed scheduled automation mechanism with a service identity, ownership metadata and alerting. The alternative problem is: A database task's schedule is stored as a local engineer's calendar reminder, so it stops running when that employee leaves. The current Timberline Foods issue is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. Its decisive acceptance evidence would be: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

The job history lists each approved target exactly once and contains no unexpected production database. That result is evidence for another corrective change: Maintain elastic job target groups through a controlled inventory and verify membership and execution scope before each rollout. The alternative problem is: A database elastic job runs on only one test database because the production target group was never updated after adding new databases. The current Timberline Foods issue is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. Its decisive acceptance evidence would be: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates. That result is evidence for another corrective change: Add approved dry-run checks, a backup or restore checkpoint and stop-on-failure conditions before the database task proceeds. The alternative problem is: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. The current Timberline Foods issue is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. Its decisive acceptance evidence would be: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

Make the task idempotent or record a durable execution checkpoint before allowing retries in the governed workflow. That relates to a separate SQL administration problem: A maintenance automation runbook retries a failed data patch and applies the same update twice because it has no idempotency check. The Granite SaaS requirement is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. That approach does not directly establish the required fix; the requested acceptance evidence is: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

Alert on per-target job execution outcomes and record failed database targets rather than using agent availability as proof of success. That relates to a separate SQL administration problem: An elastic job fails on one shard but the automation dashboard reports green because it checks only the job-agent heartbeat. The Granite SaaS requirement is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. That approach does not directly establish the required fix; the requested acceptance evidence is: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

Add approved dry-run checks, a backup or restore checkpoint and stop-on-failure conditions before the database task proceeds. That relates to a separate SQL administration problem: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. The Granite SaaS requirement is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. That approach does not directly establish the required fix; the requested acceptance evidence is: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

The task executes the intended procedure but cannot modify a restricted finance table on a target database. That result is evidence for another corrective change: Create or map a narrowly privileged job identity and grant only required procedure permissions on each approved target. The alternative problem is: A scheduled Azure SQL task retains broad database owner permissions even though it only needs to invoke one maintenance procedure. The current Granite SaaS issue is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. Its decisive acceptance evidence would be: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

An intentional retry leaves the data at the intended single-update state and the audit trail explains which attempt performed the change. That result is evidence for another corrective change: Make the task idempotent or record a durable execution checkpoint before allowing retries in the governed workflow. The alternative problem is: A maintenance automation runbook retries a failed data patch and applies the same update twice because it has no idempotency check. The current Granite SaaS issue is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. Its decisive acceptance evidence would be: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

The job history lists each approved target exactly once and contains no unexpected production database. That result is evidence for another corrective change: Maintain elastic job target groups through a controlled inventory and verify membership and execution scope before each rollout. The alternative problem is: A database elastic job runs on only one test database because the production target group was never updated after adding new databases. The current Granite SaaS issue is instead: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. Its decisive acceptance evidence would be: An investigator can reconstruct a controlled database change from job and deployment records without relying on personal messages.

Maintain elastic job target groups through a controlled inventory and verify membership and execution scope before each rollout. That relates to a separate SQL administration problem: A database elastic job runs on only one test database because the production target group was never updated after adding new databases. The Granite SaaS requirement is instead: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. That approach does not directly establish the required fix; the requested acceptance evidence is: A forced task failure reaches the active on-call channel and leaves a retrievable notification and acknowledgement record.

Record structured task identity, target resource, input version, start and end times and step results in protected logs. That relates to a separate SQL administration problem: An operations team stores only the last successful task timestamp and cannot determine which automated SQL change caused an incident. The Granite SaaS requirement is instead: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. That approach does not directly establish the required fix; the requested acceptance evidence is: A forced task failure reaches the active on-call channel and leaves a retrievable notification and acknowledgement record.

Make the task idempotent or record a durable execution checkpoint before allowing retries in the governed workflow. That relates to a separate SQL administration problem: A maintenance automation runbook retries a failed data patch and applies the same update twice because it has no idempotency check. The Granite SaaS requirement is instead: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. That approach does not directly establish the required fix; the requested acceptance evidence is: A forced task failure reaches the active on-call channel and leaves a retrievable notification and acknowledgement record.

The task runs on schedule after the engineer's account is disabled and the operations team receives completion evidence. That result is evidence for another corrective change: Move the work to a managed scheduled automation mechanism with a service identity, ownership metadata and alerting. The alternative problem is: A database task's schedule is stored as a local engineer's calendar reminder, so it stops running when that employee leaves. The current Granite SaaS issue is instead: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. Its decisive acceptance evidence would be: A forced task failure reaches the active on-call channel and leaves a retrievable notification and acknowledgement record.

A simulated invalid schema change is rejected before data mutation and a valid change passes the recovery and verification gates. That result is evidence for another corrective change: Add approved dry-run checks, a backup or restore checkpoint and stop-on-failure conditions before the database task proceeds. The alternative problem is: A patch automation workflow runs destructive schema changes without a preflight validation, health gate or recovery checkpoint. The current Granite SaaS issue is instead: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. Its decisive acceptance evidence would be: A forced task failure reaches the active on-call channel and leaves a retrievable notification and acknowledgement record.

The job history lists each approved target exactly once and contains no unexpected production database. That result is evidence for another corrective change: Maintain elastic job target groups through a controlled inventory and verify membership and execution scope before each rollout. The alternative problem is: A database elastic job runs on only one test database because the production target group was never updated after adding new databases. The current Granite SaaS issue is instead: An Azure SQL maintenance workflow sends all task notifications to a terminated employee's personal mailbox and incidents go unnoticed. Its decisive acceptance evidence would be: A forced task failure reaches the active on-call channel and leaves a retrievable notification and acknowledgement record.