Warm Queries Guide

Goal: Understand warm queries—saved queries that automatically become API endpoints.

---

What is a Warm Query?

Aspect	Description
Saved query	KQL query stored with your surface
Runs on demand	Every call executes the query against live data
API endpoint	Automatically exposed via REST
Governed	Creation requires approval, access is logged

---

Warm Query Lifecycle

CREATE → BUILD → APPROVE → DEPLOY → CONSUME
   ↓        ↓        ↓         ↓        ↓
 Empty    Azi    Human     Goes      API
 shell   writes  reviews   live    endpoint

---

Creating a Warm Query

Option A: UI

1. Enter your surface (click Surface node, then click Manage Surface icon in Inspector)
2. Drag Warm Query from the Node Bank onto the surface canvas
3. Name it (e.g., "temperature-hourly-avg")
4. Query container is created, ready for KQL

Option B: Azi

---

Building the Query

This is where Azi shines:

1. Open your warm query
2. Describe what you want in plain English:
   • "Show average temperature by hour for the last 24 hours"
   • "Find all readings above 80 degrees"
   • "Count messages by device for today"
3. Azi proposes KQL
4. Review the query
5. Approve, reject, or edit

See Working with Azi → for detailed examples.

---

Query Results

Once approved and deployed:

Access Method	How
UI preview	Click query to see results
REST API	GET /api/workspaces/explorer/warm-queries/{name}
MCP	Query via Claude, Copilot

---

Warm Query vs Ad-Hoc Query

Warm Query	Ad-Hoc Query
Saved, named	Temporary
API endpoint	No endpoint
Governed creation	Direct execution

---

Next Steps

If you want to...	Go to...
Build complex queries with Azi	Working with Azi →
Learn KQL syntax	KQL Basics →
Call queries via API	APIs Guide →