- IDENTIFY
- ANALYZE
- OPTIMIZE
- HUMAN
Short answer
How can I use AI to troubleshoot Starburst queries?
Your AI retrieves recent query history and deep execution metadata from your cluster. It identifies bottlenecks and resource issues so you can optimize your SQL. You'll see the specific technical causes in the breakdown below.
Troubleshooting outcomes.
Where your debugging leads.
The AI handles the heavy lifting of data retrieval so you can focus on the fix.
IDENTIFY
Find slow queries
The AI scans your recent history to spot queries with high latency or heavy resource usage.
ANALYZE
Inspect execution details
It pulls specific metadata to show you exactly which part of the execution plan is the bottleneck.
OPTIMIZE
Rewrite SQL logic
The AI uses the technical data to suggest better ways to structure your joins or filters.
HUMAN
Manual cluster tuning
The AI provides the data, but you still need to manually adjust cluster configurations or resource groups.
The workflow
What your AI does when a query fails or slows down.
Instead of digging through logs, you just ask your agent to look into the problem.
Scan query history
The AI looks through your recent Starburst activity to find the specific query ID you're investigating.
list_queriesPull execution stats
Once the query is identified, the AI grabs the technical breakdown of how it actually ran.
get_query_detailsPinpoint bottlenecks
The AI reviews the retrieved details to find high CPU usage or data skew issues.
get_query_detailsSuggest fixes
The AI uses the execution context to propose a more efficient version of your SQL code.
get_query_details
Try it
Copy these to start.
Use these prompts to kick off a debugging session.
Starting points
These are just ways to start the conversation. Swap out the query IDs or timeframes to match your actual situation.
Starburst Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_O3fkMMqaRDxXU8YcUOdnSQyixPMchBfro4bAKlB1/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Starburst capabilities are ready to use.
{
"mcpServers": {
"starburst-mcp": {
"url": "https://edge.vinkius.com/vk_preview_O3fkMMqaRDxXU8YcUOdnSQyixPMchBfro4bAKlB1/mcp"
}
}
}List my most recent Starburst queries so I can find the one that just failed.
Get the execution details for query ID 88291 and tell me why it's slow.
Look at the last ten queries and find any that used more than 50GB of memory.
Analyze the execution plan for query ID 4412 and suggest a better join strategy.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Starburst once, then ask.
Just provide your Starburst credentials. The connection is secure, and you'll be ready to debug immediately.
Connect Starburst to your AIFAQ
How the debugging process works
- 01
Can the AI fix my Starburst cluster settings?
No. The AI can only read query history and execution details. You have to apply any configuration changes yourself.
- 02
Can the AI delete old queries from my history?
No. This MCP only has permissions to read query data, not to modify or delete anything in Starburst.
- 03
How far back can the AI look for queries?
The AI can see whatever query history is currently available in your Starburst cluster via the list tool.
- 04
Does the AI need my Starburst password?
No. You connect through Vinkius, which handles the credentials securely so your AI client doesn't need to store them.
- 05
Can I use this to monitor real-time cluster health?
The AI can check the status of recent queries, but it isn't a real-time monitoring dashboard.
More questions about Starburst? The Connector page answers them. See everything the Starburst Connector can do
Connect Starburst to Claude, Cursor, ChatGPT & more
