Do we still need dedicated database experts with managed cloud databases?

November 17, 2026 5 answers
Asked by 7p!kettlehinge_284

We moved from on-premises databases to Aurora and Snowflake, and database work now falls to feature developers. We're seeing performance problems, questionable indexing and partitioning choices, and rising costs.

Would a dedicated database team make sense, or should developers own the data layer end to end? What should a specialist actually be responsible for?

Answers

2p+hingequartz_639

For a large enough organization, a small specialist team can be a safety net rather than the owner of every database. Feature teams still own their schemas and queries, while specialists help with difficult performance problems, design reviews, monitoring and cost anomalies.

Make the boundaries explicit. If every schema change needs a central approval, the team can become a bottleneck. If nobody knows when to ask for help, you end up with the same problems you have now. Shared patterns and review for higher-risk changes are a reasonable middle ground.

7p!kettlehinge_284:

Would query tuning and index reviews sit with the specialist team, then?

_u51~cloverfob.9:

They can help with both, especially the tricky cases. Developers should still understand and maintain the queries their features depend on.

Reply to this answer
3w+drainspindle_829

Managed infrastructure is not the same thing as a managed data model. Depending on the service and configuration, the provider handles parts of operations such as backups and upgrades. Your team still has to configure and use the service, choose schemas and storage strategies, and understand the workload.

I'd start a specialist remit with query and workload monitoring, execution plan analysis, schema reviews, indexing or partitioning choices where applicable, and cost investigation. Aurora and Snowflake are different systems, so don't assume the same tuning techniques apply to both.

Also check who owns backup configuration and recovery testing. A backup feature existing is not the same as knowing the application can be restored within the time you need.

Reply to this answer
_z8!muralcaper.2

I wouldn't jump straight to a separate team. Invest in database skills within the development teams first, because a developer who doesn't understand the data layer can build expensive problems into the application.

A dedicated full-time role may make sense at larger scale, but a team boundary can also isolate knowledge. The goal should be strong system builders, not handing all database thinking off to someone else.

Reply to this answer
q.716~claymaraca

Start with the work you need done, not the DBA title. Query and application performance, data modeling, operational maintenance and data infrastructure are related but different specialties. One person might cover several, but don't assume every database expert will be equally strong in all of them.

Look at your recurring bottlenecks and what the current team can't resolve. That should help you decide whether you need a dedicated hire, a broader team or just more depth in a particular area.

Reply to this answer
t.846~marnix

Watch the reflex to solve every slow query by making the instance bigger. Scaling can be appropriate, but it can also hide a poor query or schema while the bill keeps climbing. Someone needs enough database knowledge to tell the difference.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions