Log in
Engineering

Top 10 Database Administrator Interview Questions (2026)

Database Administrators guard one of your organization's most critical assets — the data itself. The best DBAs combine deep technical fluency with operational discipline: they prevent data loss, tune for performance before users complain, and work with developers to stop bad patterns from reaching production.

These 10 questions cover the full DBA spectrum — query optimization, high availability, backup/recovery, schema governance, and cloud-managed database operations — with guidance on what strong answers look like.

10 targeted questions Performance / HA / backup coverage 3 pro tips Updated April 2026

The 10 Interview Questions

1
A query that ran in 200ms is now taking 45 seconds. Walk me through how you diagnose and fix it.

Query performance regression is the most common DBA escalation. This scenario tests diagnostic methodology and hands-on optimization knowledge.

What to look for Strong candidates start with EXPLAIN/EXPLAIN ANALYZE (or equivalent), check for missing or unused indexes, look for statistics staleness requiring ANALYZE/UPDATE STATISTICS, investigate table bloat or lock contention, and check whether a recent schema change (new column, dropped index) or data volume increase triggered the regression. They distinguish between index seeks vs. scans, and know when a hash join vs. nested loop vs. merge join is appropriate. Look for specific tool mentions (pg_stat_statements, slow query log, wait stats on SQL Server). Weak candidates immediately suggest "add an index" without diagnosing first.
2
Describe your backup strategy for a 10TB production database with an RPO of 1 hour and an RTO of 30 minutes.

Backup strategy is the DBA's most critical reliability responsibility. This question tests whether the candidate designs for recovery time, not just backup completion.

What to look for A solid answer covers: full weekly + differential daily + transaction log every 15 minutes (to meet 1-hour RPO with margin), backup storage in a separate failure domain, regular restore tests (not just backup tests), automated restore validation, and a documented recovery runbook with verified 30-minute RTO steps. Strong candidates distinguish between logical (pg_dump) and physical (pg_basebackup, xtrabackup) backup approaches and know when each applies. They mention backup compression and retention policy. Weak candidates describe backups without ever mentioning restore testing.
3
How do you design a high-availability architecture for a database that cannot tolerate more than 30 seconds of downtime?

HA database architecture is a design and operational discipline. This question tests whether the candidate understands the full failover stack, not just replication configuration.

What to look for Look for: synchronous replication to a hot standby (eliminating data loss risk), automatic failover via a HA management tool (Patroni, Orchestrator, SQL Server Always On AG), virtual IP or DNS-based connection routing, health check intervals tuned to detect failure within seconds, and application-level connection retry logic. Strong candidates discuss split-brain prevention (fencing/STONITH), quorum requirements for automated failover, and the trade-off between synchronous replication latency and data durability. Weak candidates describe asynchronous replication as suitable for a 30-second RTO.
4
How do you manage a large schema migration (adding a column to a 500M row table) without downtime?

Zero-downtime schema changes are a critical DBA skill as tables grow. This question tests whether the candidate knows the pitfalls of naive ALTER TABLE on large tables and the techniques to avoid them.

What to look for Strong candidates describe: using pt-online-schema-change or gh-ost (for MySQL) or pg_repack (for PostgreSQL) to rebuild the table with minimal locking, batching backfills rather than running a single UPDATE on 500M rows, and validating the migration in staging with production-scale data first. They understand that a naive ALTER TABLE on a large table acquires an exclusive lock that blocks all reads and writes. Look for knowledge of database-specific DDL locking behavior and the expand/contract pattern (add column → backfill → deploy code → remove old column). Weak candidates say "schedule maintenance window" without knowing the tooling.
5
How do you enforce database access control and audit who accessed sensitive data?

Data access governance is a compliance and security requirement. This question tests whether the candidate has implemented least-privilege access control and maintains audit trails for regulated data.

What to look for Look for: role-based access control with minimal grants (application accounts with only the permissions they need), separation between read-only analytical accounts and application write accounts, row-level security for multi-tenant data, audit logging enabled for privileged operations and access to PII tables, and regular permission reviews. Strong candidates describe specific audit tooling (pgaudit, SQL Server Audit, Oracle Unified Auditing) and mention how audit logs are protected from tampering (written to a separate, immutable log store). Weak candidates describe access control without mentioning audit trails.
6
How do you approach capacity planning for a database growing at 20% per month?

Proactive capacity planning prevents performance cliffs. This question tests whether the candidate instruments and models growth rather than reacting to disk-full alerts.

What to look for Strong candidates describe: tracking key metrics (data size, index size, row growth rates by table, IOPS utilization, connection pool usage, query time trends), projecting capacity needs 3–6 months forward, and defining thresholds that trigger action before saturation. They distinguish between storage scaling (easier), IOPS scaling, and connection count scaling (different bottlenecks). They mention partitioning large tables, archiving cold data, and evaluating read replica offloading as scaling tools. Look for experience with monitoring tools (Grafana, Datadog, pgBadger). Weak candidates wait for alerts.
7
What is your approach to index design, and how do you identify indexes that should be removed?

Indexes are a double-edged sword — too few hurts reads, too many hurts writes. This question tests whether the candidate manages indexes as a portfolio, not just adds them reactively.

What to look for Look for: index creation driven by query patterns (not speculation), composite index column order chosen by selectivity and query predicates, covering indexes to eliminate table lookups for hot queries, and regular review of index usage statistics to drop unused or duplicate indexes. Strong candidates describe using pg_stat_user_indexes, sys.dm_db_index_usage_stats, or equivalent to identify bloated or dead indexes. They understand the write overhead of each index and model the trade-off. Weak candidates describe adding indexes for every WHERE clause column or never removing indexes "just in case."
8
How have you worked with development teams to improve query design before it reaches production?

The best DBAs shift performance left. This question tests whether the candidate has built developer collaboration into their workflow rather than only firefighting in production.

What to look for Strong candidates describe: code review participation for significant schema changes or new query patterns, documented query guidelines shared with developers, automated slow-query detection in staging environments (pre-production), query plan validation in CI pipelines, and office hours or pairing sessions with developers for complex data models. Look for empathy toward developers (education, not gatekeeping) and specific examples of catching N+1 queries, cartesian joins, or missing indexes before production. Weak candidates describe only reacting to production issues.
9
How do you manage database credentials and secrets in a cloud environment without hardcoding them in application configs?

Hardcoded database credentials are a persistent security vulnerability. This question tests whether the candidate has implemented modern secrets management practices.

What to look for Look for: secrets stored in a secrets manager (AWS Secrets Manager, HashiCorp Vault, GCP Secret Manager), credential rotation automated without application downtime, short-lived database credentials via Vault database dynamic secrets, IAM database authentication (RDS IAM auth), and application-level connection pooling that handles credential refresh. Strong candidates discuss the rotation strategy: how do they ensure zero-downtime rotation? They mention audit trails for credential access. Weak candidates describe environment variables as the solution without addressing rotation or audit.
10
How do you evaluate whether to use a managed cloud database service (RDS, Cloud SQL) vs. self-managed on a VM?

Cloud-managed databases have changed the DBA role significantly. This question tests whether the candidate can make pragmatic architectural recommendations rather than defaulting to either option.

What to look for Strong candidates articulate the managed vs. self-managed trade-offs: managed services reduce operational overhead (patching, backup, failover) but reduce configurability and may increase cost at high scale. They consider: team operational maturity, SLA requirements, need for custom PostgreSQL extensions, workload characteristics (managed services have instance size limits), cost at scale (self-managed is often cheaper beyond 2–4 vCPUs), and vendor lock-in. They can describe specific scenarios where each makes sense. Weak candidates default to "managed is always better" or "we always self-manage for control" without reasoning through the specific scenario.

3 Pro Tips for Hiring Database Administrators

Insights from engineering leaders who have hired DBAs across startups and enterprise teams.

Include a practical query tuning exercise

Give candidates a slow query, an EXPLAIN plan, and schema information, and ask them to identify the bottleneck and propose a fix. This practical exercise separates candidates with real production experience from those who only know DBA theory.

Test restore, not backup

Anyone can configure backups. Ask specifically how the candidate verifies that backups are recoverable, how often they run restore tests, and what the last restore drill found. A DBA who has never tested a restore under pressure is a liability waiting to happen.

Look for developer collaboration instincts

DBAs who treat developers as adversaries create bottlenecks. Ask how they have helped developers write better queries or design better schemas. The best DBAs are teachers and partners, not gatekeepers — they shift performance issues left rather than resolving them after the fact in production.

Frequently Asked Questions

What certifications are most valuable for a Database Administrator?

Oracle Certified Professional (OCP), Microsoft Certified: Azure Database Administrator Associate, and PostgreSQL certifications (pgDBA) are widely recognized. AWS Database Specialty is increasingly valuable as cloud-managed databases become the norm. Hands-on experience with production databases at scale often matters more than specific certifications.

How many interview rounds should a DBA hiring process include?

Typically 4 rounds: recruiter screen, SQL fundamentals and query optimization practical, a high availability and disaster recovery scenario discussion, and a hiring-manager behavioral round. Include a take-home schema design or query tuning exercise for senior roles.

How do you assess a DBA's disaster recovery expertise in an interview?

Present a scenario: "Your primary database server is unresponsive at 11 PM. Walk me through your response." Strong DBAs describe a methodical approach: assess failover state, initiate planned failover if available, assess data loss against last backup/replica lag, communicate to stakeholders, and document the incident. Ask for specific RTO and RPO numbers from their past environments.

What separates a great DBA from a good one?

Great DBAs are proactive, not reactive. They instrument databases to catch performance regressions before users notice, automate routine maintenance, enforce schema change governance, and partner with developers on query design before bad patterns reach production. They treat data integrity and availability as non-negotiable constraints.

Ready to hire your next Database Administrator?

Treegarden helps engineering teams structure technical interviews, collect consistent panel feedback, and make faster, fairer hiring decisions.