The most daunting screen in a DBA’s life isn’t a red dashboard; it’s the blank cursor blinking inside a query editor.

You’ve identified the bottleneck. You know the query is performing a massive table scan where it should be seeking. Now you face the tedious, manual reality of remediation: rewriting the SQL, checking syntax, verifying logic, and ensuring you haven’t introduced a regression.

In the past year, the temptation to paste a slow query into a public large language model (LLM) has grown. For a seasoned DBA, that’s an obvious nonstarter. But for a full-stack developer or a systems engineer trying to push a release, security implications aren’t always at the front of mind. They might not realize that pasting proprietary schema or customer personally identifiable information (PII) into a public chatbot is a massive compliance risk.

The results are also often erratic because the LLM lacks context. It sees the text, but it doesn’t understand the cost. That’s why we’re announcing the general availability of AI Query Assist—an intelligent tuning advisor that combines generative AI with your database execution plans. This isn’t a generic text generator. It’s a purpose-built engine designed to automate the rewrite phase of query tuning, anchored by two nonnegotiable pillars: context and security.

AI Query Assist is generally available now for:

  • SolarWinds® SQL Sentry® (SQL Server)
  • SolarWinds® Database Performance Analyzer (DPA) (SQL Server and Oracle)
  • SolarWinds® Database Observability SaaS (SQL Server)

The “Elephant in the Room”: Security and Architecture

Let’s address the immediate concern: “Where is my data going?”

At SolarWinds, we know data sovereignty is nonnegotiable. That’s why AI Query Assist is architected to ensure your database contents remain yours.

  • Sanitized Payloads: Before any data leaves your environment, AI Query Assist identifies and masks PII and sensitive literals—the service receives the structural logic of the query, not your customer data
  • Platform Connect: For our on-premises users (SQL Sentry and DPA), we utilize Platform Connect, which acts as a secure tunnel, bridging your on-prem instances to the AI service without requiring you to move the database itself or open risky firewall ports

Why do generic LLMs fail at SQL tuning? Because they treat SQL as prose and guess based on syntax patterns.

AI Query Assist is different. It not only reads the SQL text but also analyzes the Explain Plan.

By ingesting the plan, the engine understands the actual execution strategy, including cardinality estimates, join types, and index usage. This moves the output from guesswork to engineering. It allows the tool to suggest heuristics-based improvements that are mathematically likely to reduce latency, rather than just syntactically different.

How It Works: From Detection to Remediation

We’ve designed the workflow to mirror the way a typical DBA thinks, taking you from a high-wait state to a deployed fix in three steps:

  1. Select: You identify the problematic query within DPA, SQL Sentry, or SolarWinds Observability SaaS
  2. Sanitize and Analyze: Upon submission, the tool masks the data and bundles the SQL text with its specific Explain Plan
  3. Optimize: The engine generates a rewritten query focused on performance efficiency

Crucially, this isn’t a “black box.” We provide a Thinking tab that details exactly why the AI suggests a change, whether it’s converting a subquery to a JOIN, adding a missing predicate, or forcing a specific index.

Once the recommendation is ready, you can review the Original SQL and the Optimized SQL side by side.

Who Is This For?

  • For the DBA: This serves as a sanity check, handling the rote work of rewriting complex T-SQL and allowing you to focus on architectural issues—it acts as a deterministic second opinion, validating your tuning strategy against the plan data
  • For the Developer: SQL performance is a deep well, and AI Query Assist helps you learn complex patterns faster—by reading the Explanation tab, you gain insight into why your -generated query was slow and how to write it better next time

Augmented, Not Replaced

We believe in augmented intelligence. The goal of AI Query Assist isn’t to replace the DBA but to accelerate the mean time to rewrite. We handle the syntax and the initial optimization logic so you can make the final engineering decision.

Stop staring at the blank cursor.

Resources

AI Query Assist FAQ

1. What Is AI Query Assist?
AI Query Assist is a SolarWinds capability that uses generative AI plus execution plan context to suggest performance-focused SQL rewrites.

2. Why Not Use A Public AI Tool?
Public AI tools can lack execution context and may expose sensitive query details. AI Query Assist is built into your SolarWinds workflow and designed for database tuning.

3. How Does AI Query Assist Work?
You select a slow query, submit it for analysis, and review an optimized version side by side with the original, along with a clear explanation of the suggested changes.

4. Where Is AI Query Assist Available?
AI Query Assist is available for SQL Sentry, Database Performance Analyzer (DPA), and SolarWinds Observability SaaS.

5. Does It Replace Manual Query Tuning?
No. It speeds up the rewrite phase, but you still review, test, and decide whether to deploy the suggested SQL.

You may also like

Tags:
database observability
devops
sql server