Skip to content

Object Dashboard

The object dashboard provides the most detailed view of a single database object. The dashboard appears when users click a table, index, or query in the database dashboard leaderboards.

Table Detail

The table detail view displays the following metrics:

  • Table size and total size including indexes and TOAST.
  • Table bloat estimate as a percentage.
  • Sequential scan count versus index scan count.
  • Insert, update, and delete modification counts.
  • Live tuple and dead tuple counts with trend data.

Index Detail

The index detail view displays the following metrics:

  • Index size in a human-readable format.
  • Index scan count with a time-series chart.
  • Tuples read and tuples fetched counts.

Query Detail

The query detail view displays the following metrics:

  • Total and mean execution time.
  • Total rows returned and rows per call.
  • Call count with a time-series chart.

Query Plan

The query plan section appears in the query detail view below the AI Overview panel. The section defaults to expanded, and the expand/collapse state persists across browser sessions.

The panel fetches PostgreSQL EXPLAIN output when the section first renders. A refresh button in the section header regenerates the plan on demand.

Two tabs display the plan data:

  • The Visual tab shows a graphical flow diagram built from the JSON EXPLAIN plan.
  • The Text tab shows the standard EXPLAIN output in monospace format for a concise view.

Visual Diagram

The visual diagram uses a left-to-right layout. Leaf scan nodes appear on the left and the root node appears on the right. SVG bezier arrows connect each child node to its parent.

Each tile in the diagram displays the node type and the relation or index name. A colored left border indicates the cost ratio relative to the root node:

  • A red border marks nodes that consume over 80 percent of the total cost.
  • An orange border marks nodes that consume over 50 percent of the total cost.
  • The default border color applies to all other nodes.

Clicking a tile opens a popover with comprehensive node details. The popover displays the following information:

  • The cost range from startup cost to total cost.
  • The estimated row count and row width.
  • The output columns produced by the node.
  • The execution strategy and scan direction.
  • The planned and launched worker counts.
  • Any filter, join, or index conditions.

Plan Options

The system uses EXPLAIN VERBOSE for JSON plans to provide comprehensive detail in the visual mode. The text plan uses standard EXPLAIN without VERBOSE for a concise view.

For parameterized queries that use $1, $2 placeholders, the system uses the GENERIC_PLAN option available in PostgreSQL 16 and later. Older PostgreSQL versions display a friendly informational message instead of the plan.

The system caches plans for five minutes to avoid redundant queries against the database server.

AI Query Overview

The query detail view displays an AI Overview panel below the query text when an LLM provider is configured. The panel provides a brief plain-text summary of the query's performance characteristics in two to three sentences.

The AI Overview panel includes the following behaviors:

  • The summary assesses whether the query appears healthy or has potential performance issues.
  • A refresh button regenerates the summary on demand.
  • A relative timestamp shows when the summary was last generated.
  • A brain icon button opens the full AI Query Analysis dialog.
  • The system caches summaries for 30 minutes.
  • The panel is hidden when no LLM provider is configured.

AI Query Analysis

Clicking the brain icon in the AI Overview panel opens a full-screen analysis dialog. The analysis uses an agentic LLM loop with tools to gather additional context before producing a structured report.

The following tools are available to the LLM:

  • The query metrics tool retrieves historical metric values with time-based aggregation.
  • The metric baselines tool provides statistical baselines including mean, standard deviation, and extremes.
  • The database query tool executes diagnostic queries against the relevant database.
  • The schema inspection tool retrieves table and index definitions for referenced objects.
  • The query validation tool checks SQL syntax and execution plans.
  • The knowledgebase search tool finds relevant entries using similarity matching.

The dialog displays real-time progress indicators showing which tools the LLM is currently using. The final report renders as formatted markdown.

Each analysis report contains four sections:

  • The Summary section describes the query and its current performance characteristics.
  • The Performance Analysis section examines execution metrics, trends, and resource consumption.
  • The Optimization Opportunities section identifies potential improvements to the query or schema.
  • The Recommendations section suggests specific actions with supporting SQL examples.

SQL code blocks in the report include a Run button that executes the query against the correct database server. Write statements display a confirmation dialog before the system executes the query.

The system caches analysis results for 30 minutes. A Download button in the dialog footer saves the report as a markdown file.