JPJobPrepfull-stack interview
RoadmapsJS CompilerStar on GitHub
Career Roadmaps · DBA

Career roadmap

Database Administrator

Own the one thing a business genuinely cannot lose: keep it fast, correct, backed up and restorable.

Time
6-9 months part-time
Entry bar
Strong SQL and an interest in how storage engines work.
Stages
5 · 25 topics
0/25 studied0%
All roadmaps

Before you start DBA

  • SQL beyond basic SELECT
  • Linux fundamentals
  • Understanding of client/server applications

How databases actually work

5-7 weeks · 0/5 topics

Internals knowledge is what separates a DBA from a developer who writes queries.

  1. Pages, buffers and write-ahead logs explain nearly every performance behaviour.

    • Pages, heaps and row storage
    • Buffer pool and cache behaviour
    • Write-ahead logging and checkpoints
    • MVCC and version storage

  2. The single highest-value DBA skill, and the bulk of every interview.

    • B-tree structure and traversal cost
    • Composite index column order
    • Covering indexes and index-only scans
    • Partial, functional and specialised indexes

  3. Understanding why the planner chose badly is the core diagnostic skill.

    • Reading EXPLAIN ANALYZE properly
    • Statistics, histograms and cardinality estimates
    • Join algorithms and when each is chosen
    • Plan regressions and plan stability

  4. Isolation anomalies are subtle, expensive, and heavily interviewed.

    • Isolation levels and their anomalies
    • Locking, lock escalation and deadlocks
    • Long-running transactions and bloat
    • Optimistic vs pessimistic strategies

  5. Most performance problems are design problems that arrived early.

    • Normalisation and deliberate denormalisation
    • Data types and storage efficiency
    • Constraints and referential integrity
    • Partitioning strategy

BuildLoad a large dataset, then explain and demonstrate exactly how three queries are executed.

Operations and availability

5-6 weeks · 0/5 topics

The database must survive hardware failure, human error and the upgrade you postponed.

  1. Point-in-time recovery is the skill you are actually hired for.

    • Full, incremental and continuous archiving
    • Point-in-time recovery procedure
    • Restore testing and timing
    • Backup encryption and retention

  2. Synchronous versus asynchronous, and the data loss window each implies.

    • Streaming and logical replication
    • Synchronous commit trade-offs
    • Automatic failover and split-brain
    • Replica lag monitoring

  3. Major version upgrades with minimal downtime are a signature DBA deliverable.

    • Major version upgrade strategies
    • Online schema change tooling
    • Migration rehearsal and rollback
    • Cross-engine migration pitfalls

  4. Knowing which metrics predict trouble rather than merely describe it.

    • Connections, locks, replication lag
    • Slow query logging and analysis
    • Bloat, vacuum and maintenance
    • Capacity and growth forecasting

  5. The database holds the data an attacker wants. Controls here are heavily audited.

    • Roles, privileges and least privilege
    • Encryption at rest and in transit
    • Row-level security and masking
    • Audit logging

BuildConfigure replication with automatic failover, then fail over on purpose and measure the impact.

Performance engineering

5-6 weeks · 0/5 topics

The work that gets noticed: making a system faster without buying more hardware.

  1. The daily work. Interviews hand you a plan and a clock.

    • Identifying the worst queries by total time
    • Rewriting for better plans
    • Index design for a workload, not a query
    • Avoiding premature index proliferation

  2. Defaults are conservative. Knowing which knobs matter is experience made visible.

    • Memory: shared buffers, work memory
    • Checkpoint and WAL tuning
    • Autovacuum and maintenance settings
    • Connection limits and pooling

  3. Read replicas, partitioning and sharding, in the order you should try them.

    • Vertical scaling limits
    • Read replicas and read/write splitting
    • Table partitioning
    • Sharding and its operational cost

  4. Thousands of idle connections is a classic cause of mysterious outages.

    • PgBouncer and pooling modes
    • Pool sizing arithmetic
    • Connection storms and retry behaviour
    • Application-side pool configuration

  5. Most DBAs now own more than one engine, and are asked to justify each.

    Ch — MongoDB Indexing
    • MongoDB operations and index design
    • Redis persistence and eviction
    • Time-series and analytical stores
    • Choosing the right engine for a workload

BuildTake a slow production-like workload and cut its p99 latency in half, documenting every change.

Cloud and automation

4-5 weeks · 0/5 topics

Managed services changed the job. The DBAs in demand are the ones who automate.

  1. What the provider does for you, and precisely what it still does not.

    • RDS, Aurora, Cloud SQL feature differences
    • Maintenance windows and forced upgrades
    • Serverless and autoscaling databases
    • Limitations that surprise teams

  2. Provisioning, configuration and schema all under version control.

    • Terraform for database resources
    • Parameter groups as code
    • Migration tooling in pipelines
    • Environment parity

  3. The DBA who blocks deployments becomes a bottleneck; the one who automates gates does not.

    • Migration review automation
    • Backwards-compatible schema changes
    • Blue/green database deploys
    • Rollback strategy for data changes

  4. Turning your own expertise into checks that run without you.

    • Automated slow query reporting
    • Index usage and bloat reports
    • Anomaly detection on key metrics
    • Self-service dashboards for developers

  5. Databases are usually the largest line on a cloud bill.

    • Instance right-sizing from real metrics
    • Storage tiering and archival
    • Reserved capacity planning
    • Query cost as an optimisation target

BuildProvision a replicated database entirely from code, with automated backup verification.

Interview preparation

3-5 weeks · 0/5 topics

DBA interviews are practical: a plan, a lock, a failed restore, and how you reason.

  1. Advanced SQL under time pressure, usually including window functions.

    • Complex joins and subqueries
    • Window functions and ranking
    • Set operations and CTEs
    • Writing queries that use an index

  2. Here is a slow query and its plan. Talk me through it.

    • Reading a plan out loud
    • Proposing index changes with reasons
    • Estimating the improvement
    • Knowing when the fix is in the application

  3. Recovery and incident scenarios dominate senior DBA interviews.

    • Restore to a point before a bad deploy
    • Replication broken, what now
    • Disk nearly full on the primary
    • Deadlocks appearing after a release

  4. Design a schema and a scaling plan for a described workload.

    • Schema for a given access pattern
    • Partitioning and archival strategy
    • Read scaling and caching
    • Choosing the engine with justification

  5. You are being trusted with the company's data. Judgement stories matter.

    • A recovery you performed under pressure
    • A migration you ran with no downtime
    • Refusing an unsafe change
    • Teaching developers to write better queries

BuildA written portfolio: a tuning case study, a failover test report, and a recovery runbook.

DBA tools on your CV

  • PostgreSQL
  • MySQL
  • SQL Server
  • MongoDB
  • PgBouncer
  • Terraform
  • pgBadger
  • Prometheus

What DBA employers ask to see

  • A tuning case study with EXPLAIN plans and latency numbers
  • A tested point-in-time recovery with timings
  • A failover drill report
  • An automated database provisioning repository

Managed cloud databases removed some routine work and increased demand for people who can actually diagnose and design. Finance, healthcare and any high-volume transactional business hire steadily.

Content last reviewed 2026-08-31. Guidance only — no institute or paid placement is endorsed anywhere in this book.