The first Friday in July is the designated DBA Appreciation Day every year. Seems fair, right? You manage the most important resource an organisation has, its data, and you get one day in early July to be celebrated.

This year, DBAs deserve a little extra appreciation because it’s so easy to forget what makes the role important in a world of DevOps and encroaching AI/ML tooling.

Are DBAs still relevant?

The line between developer and system administrator gets a little fussier every year. The rise of DevOps, SREs, demand for ‘full stack developers,’ not to mention Database-as-a-Service (DBaas) offerings and the general push to the cloud, calls into question specialisation.

Add to that the explosion of AI/ML tools that will happily give advice on SQL queries and database configurations, we have to ask whether database administrator is still a role we need – or what the future holds for DBAs.

The answer is a resounding yes

The answer, of course, is a resounding yes. The landscape is changing, and others are expected to know a bit more about database administration than years past, but DBAs are still in high demand. Like other IT admin roles, though, there’s less focus on grunt work and more focus on higher-level skills and work.

For example, provisioning databases is something that should be automated in any large organisation. Developers should be able to request and receive a new database in minutes, not hours, days or (ugh) weeks.

But DBAs should be responsible for the overall system, ensuring that it meets security, data retention, and resource policies.

Developers should be able to write queries, but somebody needs to be able to help them tune queries and observe database performance. Organisations need DBAs to help plan things like MySQL 5.7 to 8.0 migrations, to ensure that the organisation has adequately planned for disaster recovery.

Is your organisation using a DBaaS product? Your DBA should be helping to ensure that you’re not over-spending and that you have a strategy to avoid lock-in if your organisation decides to move to another solution in the future.

What about AI/ML?

It’s true that ChatGPT and other tools can give you answers to simple and complex questions about things like SQL queries and database tuning. It will confidently answer questions like, “how large should my innodb_redo_log_capacity setting be for MySQL if my application writes 100MB every hour?”

It will also stand a good chance of being very wrong. The more complex the, pardon the pun, query posed to ChatGPT or another AI/ML tool – the greater the chance that it will fabricate an answer that is wholly wrong (note that I didn’t use the industry-preferred term ‘hallucinate.’ Let’s call a fabrication a fabrication.)

No doubt, AI/ML tools will continue to improve and someday they may be much more reliable. But they’re no substitute for a DBA who has deep expertise with not only the database in question but the environment it’s running in and the business requirements that it needs to meet.

The skills that matter

It’s true that some traditional DBA work and skills have been absorbed into other roles, or obsoleted by more modern tooling. If it can be automated or done by generalists, it should be. This gives DBAs the opportunity to focus on work that is more strategic, adds more value, and cuts down on firedrills.

Today’s DBAs will be focusing on migrating from on-prem to cloud environments, not to mention figuring out data strategies that mesh with cloud-native apps on Kubernetes.

Just as other players in the organisation have had to learn some DBA skills, DBAs are going to have to broaden their skillsets too. It may not be enough to have amazing PostgreSQL skills, you’ll probably need to learn your way around AWS, Azure, or other public cloud platforms.

Companies need DBAs who can inform the company’s data strategy, ensure that the data is secure and access to sensitive data is tightly controlled. DBAs who can inform the organisations’ monitoring strategies and implement sophisticated and predictive monitoring and telemetry to help optimise performance.

All of that remains very necessary, very relevant, and should be appreciated year-round.