DP-300 question patterns
Find a pattern
Each row is wording that keeps showing up in questions, with the answer it usually points to in green. Tap a row to see why.
In this list
No pattern matches that. Try a service name, such as CodeDeploy or Config.
Wording that decides the answer
-
SQL Managed Instance.
It has near-complete instance-level compatibility. SQL Database lacks Agent and cross-database queries.
-
SQL Server on an Azure VM.
Only IaaS gives OS control and version choice.
-
Prefer SQL Database or Managed Instance over a VM.
Patching, backups and HA are automatic in PaaS.
-
An elastic pool.
Databases share a resource budget, so cost is lower than sizing each for peak.
-
Serverless vCore compute with auto-pause.
Compute is paused when idle; storage is still billed.
Planning and migration
-
Azure Arc-enabled SQL Server.
It projects servers into Azure for inventory, security, patching and billing.
-
SQL database in Microsoft Fabric.
It mirrors automatically to OneLake in Delta format.
-
Table partitioning with a date partition function and `SWITCH`.
Switching is a metadata operation.
-
Elastic database tools with a shard map manager.
It routes requests by sharding key.
-
Log Replay Service, DMS online mode or the Managed Instance link.
They continuously apply log backups until cutover.
-
`CREATE DATABASE ... AS COPY OF`.
Online, transactionally consistent copy.
-
Managed Instance database copy and move.
It avoids backup and restore plumbing.
Security
-
Enable Microsoft Entra-only authentication.
SQL logins are disabled at the server.
-
Managed identity with `CREATE USER ... FROM EXTERNAL PROVIDER`.
No credential to leak or rotate.
-
Always Encrypted.
Keys never reach the engine.
-
Always Encrypted with secure enclaves.
The enclave computes on plaintext inside a protected region.
-
Transparent data encryption.
It encrypts files at rest, not data in use.
-
TDE protector in Azure Key Vault.
The customer controls and can revoke the key.
-
Dynamic data masking.
It masks results without changing storage.
-
Security policy with a filter predicate.
The engine enforces it for every query.
-
Ledger tables with digests in immutable storage.
Digests outside the database detect tampering.
-
Private endpoint with private DNS.
Service endpoints aren't reachable from on-premises.
-
Server-level auditing to Log Analytics.
It covers every database on the logical server.
-
Change data capture.
Change tracking only records that a row changed.
Monitoring and performance
-
Query Store, then force the older plan.
It keeps plans and runtime stats.
-
Database watcher with a Data Explorer store.
It replaced SQL insights.
-
Intelligent Insights diagnostics log.
It gives root cause; not on MI.
-
An Extended Events session with an event file target.
It replaced Profiler and has low overhead.
-
Query `sys.dm_exec_requests` and follow `blocking_session_id`.
The head blocker is the root of the chain.
-
Enable FORCE_LAST_GOOD_PLAN.
It reverts to the last good plan and validates.
-
Enable CREATE_INDEX and DROP_INDEX.
Not available on MI or VMs.
-
`ALTER INDEX ... REBUILD`.
Above 30% rebuild; 5-30% reorganize.
-
Resource pool, workload group and classifier.
Not available in SQL Database.
-
Raise compatibility level for intelligent query processing.
Feedback adjusts grants on later runs.
Automation
-
An elastic job, or Azure Automation or Logic Apps.
SQL Database has no Agent.
-
Elastic job with a server or pool target group.
Targets are re-enumerated each run.
-
Bicep template deployed in a pipeline.
It is declarative and idempotent.
-
`what-if` on the deployment.
Shows adds, changes and deletes without applying.
-
Metric alert, action group and an Automation runbook.
Automation changes the service objective.
High availability and recovery
-
Auto-failover group with the read-write listener.
Listener names stay the same.
-
Active geo-replication.
Per-database, up to four secondaries.
-
Always On availability group.
Replicates databases; WSFC plus listener.
-
Failover cluster instance with shared disks or Premium file share.
Whole instance fails over; one copy of data.
-
Log shipping.
Backs up, copies and restores logs on a schedule.
-
Point-in-time restore to a new database.
PaaS restores never overwrite.
-
Long-term retention policy with a yearly retention.
LTR holds full backups up to 10 years.
-
Restore with NORECOVERY until the last step.
RECOVERY ends the chain.
-
Managed Instance link (distributed AG).
Near-real-time replication to MI.
-
Planned failover drills and test restores with measured RTO and RPO.
Untested plans fail.
Now spot them in real questions. Practice sets use the same wording.
Practise 20 questions