Microsoft Fabric RTI Expert

Expert guidance for Microsoft Fabric Real-Time Intelligence using the Fabric RTI MCP Server. Work with Eventhouse databases, Eventstreams, Activators, and Maps through natural language.

Core Capabilities

1. Eventhouse/KQL (12 tools) - Query data, manage schemas, sample data
2. Eventstreams (17 tools) - Build real-time streaming pipelines
3. Activator (2 tools) - Create triggers and alerts
4. Map (7 tools) - Manage data visualizations

Quick Reference

Eventhouse Tools

| Tool | Purpose | |------|---------| | kusto_query | Execute KQL queries | | kusto_list_tables | List tables in database | | kusto_get_table_schema | Get table schema | | kusto_sample_table_data | Sample table records | | kusto_ingest_inline_into_table | Ingest CSV data |

Eventstream Tools

| Tool | Purpose | |------|---------| | eventstream_list | List Eventstreams | | eventstream_create | Create new Eventstream | | eventstream_add_sample_data_source | Add sample data source | | eventstream_add_eventhouse_destination | Add Eventhouse destination | | eventstream_validate_definition | Validate configuration |

Activator Tools

| Tool | Purpose | |------|---------| | activator_list_artifacts | List triggers | | activator_create_trigger | Create alert trigger |

Map Tools

| Tool | Purpose | |------|---------| | map_list | List Maps | | map_create | Create new Map | | map_get_definition | Get Map configuration |

Querying Data with KQL

kusto_query

Execute KQL queries against Eventhouse databases.

Parameters:

• database (string) - Target database name
• query (string) - KQL query text

Example:

StormEvents
| where State == "ILLINOIS" and EventType == "Flood"
| summarize Count=count() by StartTime
| order by StartTime desc

kusto_sample_table_data

Get sample records from a table.

Parameters:

• table_name (string)
• sample_count (number, default: 10)

Managing Eventstreams

eventstream_create

Create a new Eventstream for real-time data processing.

Parameters:

• workspace_id (string)
• display_name (string)
• description (string, optional)

eventstream_add_eventhouse_destination

Route data to Eventhouse for analytics.

Parameters:

• eventhouse_id (string)
• kql_database_id (string)
• table_name (string)
• input_serialization_type (string) - "Json", "Csv", etc.

Workflow

1. eventstream_start_definition
2. eventstream_add_sample_data_source
3. eventstream_add_eventhouse_destination
4. eventstream_validate_definition
5. eventstream_create_from_definition

Creating Activator Triggers

activator_create_trigger

Create triggers for real-time alerting.

Parameters:

• workspace_id (string)
• display_name (string)
• description (string)
• eventhouse_id (string)
• kql_database_id (string)
• query (string) - KQL query for monitoring
• notification_type (string) - "Email", "Teams"
• recipients (array) - Email addresses or Teams webhooks

Example: Monitor for floods and send email alert

Query: StormEvents | where EventType == "Flood" and State == "ILLINOIS"
Notification: Email to admin@company.com

Common Scenarios

Query Analysis

1. kusto_list_databases - Find databases
2. kusto_list_tables - Find tables
3. kusto_get_table_schema - Understand structure
4. kusto_query - Run analysis query

Real-Time Pipeline

1. eventstream_create - Create pipeline
2. eventstream_add_custom_endpoint_source - Add data source
3. eventstream_add_derived_stream - Transform data
4. eventstream_add_eventhouse_destination - Save to database
5. eventstream_validate_definition - Check config

Alerting Setup

1. kusto_query - Test alert condition
2. activator_create_trigger - Create alert
3. Monitor for notifications

When to Use This Skill

• Querying Fabric Eventhouse with KQL
• Building real-time data streaming pipelines
• Creating data-driven alerts and triggers
• Managing real-time analytics workloads
• Working with time-series and event data
• Implementing event-driven architectures

Keywords

microsoft fabric, real-time intelligence, rti, eventhouse, kql, kusto, eventstream, activator, map, real-time analytics, streaming data, event-driven, triggers, alerts, time-series data