Direct database MCP connector
Connects an AI client to one configured database so it can inspect schema, read data or perform approved operations. Its view is normally limited to that connection and the evidence available when the question is asked.
Mini DBA provides one multi-database MCP Server that lets compatible AI clients investigate performance evidence already collected by Mini DBA Engines. Ask across a mixed SQL Server, PostgreSQL, MySQL, MariaDB, Oracle and Azure SQL estate instead of configuring a separate investigation workflow for every platform and server.
- Explore live and historical monitoring data through Model Context Protocol tools.
- Ask across SQL Server, PostgreSQL, MySQL, MariaDB, Oracle and Azure SQL estates.
- Drill from an estate-wide symptom into affected servers, workloads and time ranges.
- Use API keys to control access from approved AI clients.
Monitoring MCP versus direct database MCP access
Both approaches can be useful, but they solve different problems. Mini DBA is designed for operational investigation across monitored systems rather than unrestricted AI-generated database access.
Mini DBA monitoring MCP
Gives approved AI clients access to live and historical monitoring data from connected Engines. It can start with an estate-wide symptom, find the affected platform and server, then retrieve the metrics, waits, queries, alerts and incident history that explain it.
Use them for the right job
Use direct connectors when an agent needs controlled application-data operations. Use Mini DBA when DBAs, developers or support teams need performance triage, incident reconstruction and risk discovery across a mixed estate.
Ask operational questions across the estate
A connected AI client can combine several Mini DBA tools in one investigation, moving from a broad question to the evidence behind it.
Live problem drilldown
Ask which servers show the strongest signs of pressure, then investigate sessions, blocking, I/O, waits, queries, alerts or execution plans for the affected systems.
Likely future problems
Ask about worsening trends, repeated alerts, capacity pressure or recurring workload patterns that deserve attention before they become a service incident.
Historical investigation
Use collected history to explore when a problem started, what changed around the event and whether the same signature has appeared elsewhere.
Powerful for support teams and managed service providers
Support companies and MSPs can ask consistent questions across monitored client estates while keeping each investigation grounded in Mini DBA evidence.
Across multiple clients
- Find clients with active critical alerts or disconnected servers.
- Identify estates showing repeated blocking, resource saturation or capacity risk.
- Compare similar symptoms across separate client environments.
- Build an evidence-led starting point for daily reviews and incident triage.
During a client incident
- Move from the affected client to the server, database and workload involved.
- Ask for a concise timeline using metrics, alerts and available history.
- Surface the likely next diagnostic checks for an engineer to verify.
- Create a clear technical handover without losing the original evidence.
Explore the complete Mini DBA database monitoring workflow for MSPs and support companies.
Ask one question, then follow the evidence across platforms
The MCP client does not have to know which database platform is responsible before an investigation begins. Start at estate level, identify the affected server, then use platform-specific evidence for the next step.
1. Scope the incident
Ask which Engines, sites, servers or databases show active pressure, unusual workload, alerts or connection problems in the relevant time window.
2. Drill into the platform
Follow SQL Server waits and plans, PostgreSQL locks and vacuum evidence, MySQL or MariaDB InnoDB activity, Oracle sessions and waits, or Azure SQL resource pressure.
3. Compare with history
Check whether the symptom is new, recurring or growing, then use the Console to verify the evidence and retain normal human review and change control.
Connect AI to your database monitoring estate
Read the MCP Server setup guide or explore the in-page AI Assistant.
One Database MCP Server across multiple database platforms
Use the same AI-assisted investigation workflow across a mixed estate while retaining the platform-specific metrics, queries, waits, alerts and history collected by Mini DBA.
SQL Server MCP monitoring
Investigate blocking, waits, deadlocks, expensive queries, execution plans, memory, I/O and recurring SQL Server performance problems.
PostgreSQL MCP monitoring
Ask about slow queries, long transactions, lock contention, vacuum pressure, WAL activity and PostgreSQL capacity trends.
Oracle MCP monitoring
Explore Oracle sessions, waits, SQL performance, memory, I/O, tablespace capacity and historical database health.
MySQL and MariaDB MCP monitoring
Review query load, connections, InnoDB pressure, locks, deadlocks and replication evidence with an approved AI client.
Azure SQL MCP monitoring
Find the queries and waits behind DTU, vCore, I/O, transaction-log and elastic-pool pressure in Azure SQL.
Explore the complete cross-platform database monitoring workflow behind the MCP tools.
Use your preferred MCP-compatible AI client
Copy the endpoint from Mini DBA MCP Server settings into a compatible local or managed AI client, then authenticate with a dedicated API key. Your client discovers the database monitoring tools exposed by Mini DBA and can call the relevant tools as you ask questions.
This makes Mini DBA monitoring data available in the AI workflow your engineers already use, subject to that client's MCP support, network access and security policy.
Example questions
- Which monitored servers need attention right now?
- What is causing the current blocking on this client estate?
- Which systems show a growing memory or I/O risk?
- Have these alerts or query symptoms happened before?
- Which execution plans should an engineer review first?
Designed for controlled access
Dedicated API keys
Create separate keys for individual clients or integrations so access can be rotated or removed without disrupting every MCP connection.
Your network boundary
Choose which trusted AI clients can reach the MCP endpoint. Monitoring data can include server names, database names, query text and operational history.
Read and verify
Use MCP answers to guide investigation and verify important conclusions against Mini DBA screens, operational policy and normal change control.
Community and licensed MCP access
Community Edition can expose its available live and historical monitoring tools through MCP, so you can ask an approved AI client about the server it monitors. Community Engines do not evaluate or retain alerts, so Community alert records are not exposed through MCP.
Licensed Engines add their licensed alert data and can monitor the number of database connections included by the licence. In a mixed Console, entitlement is applied by the Engine that owns each monitored server.
Database MCP Server FAQ
What is a database MCP server?
It provides compatible AI clients with approved tools through the Model Context Protocol. Mini DBA exposes monitoring evidence already collected by its Engines rather than making you assemble metrics and history manually.
Can it investigate multiple servers?
Yes. Ask across the servers and database platforms available through the Console, including optional multi-Engine, multi-site and managed-client environments.
Does it replace the Mini DBA Console?
No. MCP provides another investigation route. Use the AI response to guide the investigation, then verify important conclusions against Mini DBA screens and your normal operational controls.
Does the AI client need every database password?
No. The client authenticates to the Mini DBA MCP endpoint. Mini DBA Engines remain responsible for their configured monitoring connections, so individual database credentials do not need to be copied into the AI client.
Investigate the database problem, not just the platform
Use a problem-led workflow across SQL Server, PostgreSQL, MySQL, MariaDB, Oracle and Azure SQL.
Blocking and lock contention
Find the lead blocker and understand every waiting session.
Investigate blocking →Deadlock analysis
Preserve participants, statements and resources after the event.
Analyse deadlocks →Query performance
Prioritise expensive workload and inspect execution plans.
Analyse queries →Performance history
Reconstruct incidents after the live symptoms have gone.
Explore history →Evaluate Mini DBA with confidence
Review how monitoring data and access are handled, then compare the available ways to use Mini DBA.
Security and data handling
Understand deployment boundaries, permissions, credentials, AI context, MCP access and external integrations.
Review Mini DBA security →Compare monitoring approaches
Compare Community, trial and licensed use, and see where continuous monitoring differs from scripts and snapshots.
Compare Mini DBA →