According to recent data from the SolarWinds “State of Monitoring and Observability Report,” 51% of IT pros believe database performance would benefit from better observability. Oftentimes, database observability suffers because teams have multiple, disparate observability and monitoring tools. This tool sprawl creates blind spots that make it harder to identify and react to database performance issues.
As with many IT workflows, AI has the potential to streamline database monitoring and reduce the manual load traditionally associated with certain observability workflows. Although AI scales monitoring and observability for modern environments, there are still certain mistakes database administrators (DBAs) make when it comes to using AI.
It is important to identify those mistakes, address them, and implement robust AI and observability best practices to improve database performance.
Where AI Shows Up in Database Management
Currently, there are two main ways AI appears in DBA functions. First, generative AI (GenAI) streamlines many manual tasks, including bridging knowledge gaps, synthesizing large sets of information, and writing documents. The subject matter for a DBA can be vast, so GenAI helps keep their bases covered and fills those gaps.
Secondly, GenAI is becoming an asset for performance improvement. The most common example is in SQL code generation. The technology can suggest alternatives, test code, debug, and refactor. In addition, AI streamlines many aspects of database design, such as designing indexes to simplify database search and improve overall performance. GenAI helps DBAs improve performance when it is used to analyze and interpret performance data.
This includes correlating the occurrence of specific error messages to performance telemetry, which provides the DBA with a solid basis to move forward when remediating.
DBA Shortfalls of AI in Database Management
Even with GenAI’s current role in database management, there are still hurdles that can get in the way of valuable GenAI use. The role of a database administrator has become a paradox. DBAs may sit between a requirement for demanding and stressful work, while battling the requirement to work fast. Conversely, AI can both relieve workload and streamline workflows. It’s important for DBAs to understand how GenAI works as well as the nuances of the database platforms GenAI supports.
With the pressures to work quickly, which I like to call “the tyranny of the urgent,” DBAs might take outputs from GenAI and directly apply its recommendations with little to no verification. This includes practices such as implementing full SQL script recommendations without careful review. Furthermore, even if reviews exist, DBAs may not take steps to understand the logic behind the output. AI shortfalls can also appear before use or during implementation.
The aforementioned DBA need for speed may contribute to fast, ad hoc AI tooling deployments. This method of AI implementation can later have negative consequences on both the DBA and the database they manage. To avoid these consequences, there are at least two specific considerations organizations should make before implementing AI:
- Compliance and explainability—There is a growing aversion to the black box nature of AI in many states and regions. As a result, sovereign states around the world are demanding that AI tooling be more transparent in its decision making. One example is the 2026 EU AI Act. Many of these new laws and regulations are expected to contain significant penalties for breaches.
- Human rights—Similar to explainability, there are a growing number of regulations around the world that are expanding the protection of human rights in a digital world. For example, the European Union’s General Data Protection Regulation includes a “right to be forgotten” clause, which gives people the right to have their personal data erased from AI models and corporate databases, among other systems. Erasure can be easy in a relational database, as DBAs can simply delete personal data within a given database. However, GenAI commonly stores data using alternative techniques—such as vector database and modeling weights—which are much more opaque and difficult to scrub.
Best Practices and Proper AI Tooling for Database Observability
After properly considering these factors, teams can slowly and steadily implement AI observability tooling into their database management workflows. The first step is to identify which tasks are the most time-consuming for DBAs. Whether it’s responding to tickets, anomaly detection, or event log reviews, there are AI-powered observability solutions that can dramatically reduce the need to spend all day on these tasks. With an AI-powered unified monitoring and observability platform, teams can automate these workflows while simultaneously monitoring entire databases through a single pane of glass.
Making AI a true value add requires teams to do more than select the right tooling. They must also do the following:
- Continue to tune alerts for error monitoring and resource consumption thresholds to avoid alert fatigue and prioritize actionable alerts.
- Prioritize and set automated responses to each alert. For example, don’t just raise an alarm that a database backup process has failed. Instead, create a workflow for failed database backups that automatically initiates a new backup, according to relevant workflows, whenever a critical backup fails.
- Relentlessly review the information that comes from database monitoring. This will allow teams to perceive trends in database activity and find opportunities to increase efficiency. It will also allow DBAs to enact human-in-the-loop practices when AI solutions deliver a finding provided by their monitoring insights.
How AI Will Evolve Database Management
AI will soon permeate every aspect of database management, evolving from an option in the DBA’s toolbelt to a trusted advisor for each workflow. DBAs will increasingly implement more AI tooling to gather information across disparate data sources and inform database strategy. Take, for example, budgeting and cost optimization for cloud computing database environments.
The complexity and volume of information related to cost optimization can make any DBA lose sleep. But with GenAI, financial operations and capacity planning can be performed quickly, easily, and with minimal anxiety.
IT professionals, including DBAs, will also shift from operators to orchestrators. In fact, according to recent data, 80% of IT professionals already anticipate this shift. The DBAs of today are doers. Think of them like the mechanics on a Formula 1 racing team. Future GenAI capabilities will transform DBAs from a mechanic who fixes the database engine to the race engineering manager who optimizes the vehicle and the team for winning the race. The former is someone who gets their hands dirty by going deep into the internals of the machine; the latter has several team members (i.e., AI agents), each responsible for a different part of bringing home the win. The DBAs of the 2030s will be conductors of an agentic orchestra.
To reach this future, however, enterprises must partner with their database teams to create a strong foundation for deeper AI implementation as soon as possible. This means providing DBAs with the right AI-powered observability tooling, enacting a methodical and responsible approach to AI implementation, and integrating AI with a unified, enterprise-wide monitoring system.