Live Queries

The Live Queries page is the primary view in IndeX-Ray. It connects to your database through the analyzer, pulls in every query tracked by pg_stat_statements, and runs each one through an optimization analysis — surfacing index recommendations, cost reductions, and augmented explain plans.

The query list

The main view is a sortable table of all captured queries. Each row shows:

Column	Description
Query	Cleaned SQL text (first two lines)
Status	Optimization state (see below)
Cost	Estimated cost reduction (baseline to optimized), if available
Nudges	Count of suggestions grouped by severity (critical, warning, info)
First Seen	When the query first appeared in pg_stat_statements
Mean Time	Average execution time in milliseconds
Calls	Total execution count
% of Runtime	Share of total database runtime

Click any column header to sort. The default sort ranks queries by nudge severity so the most actionable queries surface first.

Queries sync from the database every 60 seconds automatically. Use the Refresh button to sync immediately.

Filtering

The left sidebar provides filters to narrow the list:

• Affecting Table — show only queries that touch a specific table
• Only SELECT queries — hide writes (INSERT, UPDATE, DELETE)
• Filter out system queries — hide queries against pg_* and information_schema
• Filter out targetless queries — hide SELECTs without a FROM clause (e.g. SELECT 1)
• Only improved queries — show only queries where optimization found a better plan
• Status — multi-select to show/hide specific optimization states

All filter preferences persist in your browser across sessions.

Optimization states

Each query goes through an optimization pipeline. The status column reflects its current state:

Status	Meaning
Waiting	Queued for analysis
Optimizing	Currently being analyzed
Improvements available	Found index recommendations with a cost reduction
No improvement found	Query is already optimal given the current schema
Not supported	Query type can’t be optimized (e.g. DDL, utility)
Timeout	Analysis exceeded the time limit
Error	Analysis failed

The status bar at the bottom of the page shows overall progress: how many queries have been processed, how many have improvements, and how many are still waiting.

Query detail

Clicking a query opens a detail view with four sections:

Metadata

Top-left panel showing the query’s key metrics: status, cost, mean time, calls, runtime share, and first-seen date.

Insights

Top-right panel containing:

• Nudges — optimization suggestions grouped by severity (critical, warning, info), with full explanations
• Suggested indexes — concrete CREATE INDEX recommendations when improvements are available
• Tables accessed — every table the query touches
• Indexes used — indexes referenced in the current plan, linked to the Indexes page
• Tags — SQLCommenter metadata (source file, route, method) if present

Formatted query

Bottom-left panel with the SQL text, annotated inline with nudge markers showing exactly where in the query each suggestion applies.

Explain plan

Bottom-right panel with the query’s execution plan. When improvements are available, you can toggle between the Baseline plan (current schema) and the Optimized plan (with suggested indexes applied) to see the difference.

Three view modes are available:

• Tree — interactive visual tree with expandable nodes, showing augmented fields like cost breakdowns, alternative plans, ineligible indexes, and order analysis
• ASCII — text-based tree for quick scanning or copy-pasting
• JSON — raw explain plan output

Use the previous/next arrows (or keyboard arrows) in the detail header to navigate through your filtered query list without returning to the table.

Toolbar actions

• Refresh — manually trigger a sync from the database
• Export to JSON — download all queries with improvements as a JSON file
• Reset queries — clear pg_stat_statements on the source database (see source code mapping for when this is useful)
• Settings — configure your IDE preference for clickable file links