Top 6 Database Administrator Interview Questions (2026)

Behavioral questions

Past-experience questions. Answer with the STAR method: Situation, Task, Action, Result.

1. 1

   Tell me about a database performance or availability problem you diagnosed and resolved under pressure.

   What they're really asking: Production incident experience: diagnosis under time pressure, communication with the business while working the problem, the actual fix, and the post-incident change to prevent recurrence. Database incidents are high-visibility and the story reveals how the candidate performs when it matters.

Technical questions

Skill and knowledge checks. Be specific — name tools, tolerances, and methods.

1. 1

   A query that used to run in 2 seconds is now taking 45 seconds. Walk me through your optimization approach.

   What they're really asking: Query optimization methodology: check the execution plan first, look for table scans where index seeks are expected, check index fragmentation, verify statistics are current, check for blocking or locking, and look at whether the data volume has grown past a threshold the current plan handles well.

   Strong answer (structured troubleshooting):

   Execution plan first

   I pull the actual execution plan, not the estimated one. The difference between estimated and actual row counts is often where the problem lives — the optimizer built a plan based on outdated statistics, and the plan is catastrophically wrong for the actual data.

   Look for scans

   I look for table scans or index scans where seeks are expected. A scan on a large table is often the symptom; a missing index or a non-sargable predicate (functions on the indexed column in the WHERE clause) is usually the cause.

   Statistics and fragmentation

   If the plan looks correct but the query is still slow, I check when statistics were last updated and whether index fragmentation is above the threshold that hurts scan performance. On a growing table, both go stale faster than the maintenance job runs.

   Blocking

   If it's intermittently slow rather than consistently slow, I check for blocking with sys.dm_exec_requests and sys.dm_os_waiting_tasks. A query that's waiting on a lock looks slow but isn't a query problem — it's a locking problem.

   Starting with the actual execution plan — not guessing or adding indexes blindly — is the discipline that makes a DBA efficient. Every other step follows from what the plan shows.

   Practice answering this question out loud →
2. 2

   Explain the difference between clustered and non-clustered indexes.

   What they're really asking: Index fundamentals: a clustered index determines the physical order of rows in the table (one per table, usually the primary key); non-clustered indexes are separate structures that point back to the clustered index or heap. The clustered index is the table; non-clustered indexes are lookups into it.

3. 3

   Describe your backup and recovery strategy for a production database.

   What they're really asking: RPO and RTO thinking: full backups, differential backups, transaction log backups (for point-in-time recovery), backup verification (a backup you haven't tested is not a backup), offsite or cloud backup storage, and documented recovery procedures that someone other than you can execute at 2 AM.

4. 4

   What is a deadlock and how do you resolve and prevent it?

   What they're really asking: Deadlock fundamentals: two transactions each holding a lock the other needs, waiting indefinitely. SQL Server resolves them automatically by killing the victim transaction. Prevention: consistent lock acquisition order across transactions, short transactions, appropriate isolation levels, and row-level versus table-level locking.

5. 5

   How do you approach database security — who should have access to what?

   What they're really asking: Least privilege and role-based access: applications connect with service accounts that have only the permissions they need (execute stored procedures, not direct table access), DBAs have elevated access with audited logins, and no production access through shared accounts.

How to prepare for a Database Administrator interview

• 1

  Execution plans are the job

  Every SQL Server or PostgreSQL DBA interview will eventually get to query performance. Practice reading actual execution plans — identifying the expensive operators, understanding why the optimizer chose a plan, and knowing how to influence the plan without hints.

• 2

  Know your platform's HA options

  SQL Server: Always On Availability Groups, failover clustering. PostgreSQL: streaming replication, logical replication. The specific HA architecture you've operated — not just read about — is what interviewers want to hear about.

• 3

  Backup testing is the credibility signal

  DBAs who can describe their last successful restore test — including how long it took — are the ones who've actually validated their backups. 'We have backups' is not a recovery plan.

• 4

  Ask about their monitoring and maintenance window situation

  Whether they have automated monitoring and alerting, how maintenance windows are scheduled, and what their RTO/RPO targets are tells you about the operational maturity and the on-call expectations.

Database administrators remain in steady demand across enterprises of all sizes, with cloud database experience (AWS RDS, Azure SQL, Google Cloud SQL) growing as a differentiator alongside traditional on-premises expertise. DBAs who combine performance tuning skills with cloud platform knowledge and DevOps database practices (CI/CD for schema changes, infrastructure as code) are particularly sought after.