TiDB Dashboard

The TiDB Dashboard is a web UI for managing and monitoring TiDB clusters.

Accessing the Dashboard

http://<pd-ip>:2379/dashboard

• Username: root
• Password: (empty by default)

Key Features

1. Cluster Information

Displays basic cluster information:

• Number of TiDB, TiKV, TiFlash, and PD nodes
• Status of each node (online/offline)
• TiDB version
• Disk capacity and usage

2. Key Visualizer

Displays data distribution across TiKV nodes:

• Region distribution — data heatmap visualization
• Hot regions — frequently accessed Regions
• Read/Write traffic — read/write traffic

3. SQL Statements

Displays the most frequently executed SQL statements:

• Duration — average execution time
• QPS — queries per second
• Scan Rows — number of rows scanned
• Memory — memory usage

4. Slow Queries

Displays slow-running queries:

• Query detail — the slow SQL statement
• Execution plan — execution plan analysis
• Duration breakdown — time components

5. Key Performance Metrics

The dashboard displays key metrics:

• QPS — total queries per second
• TPS — total transactions per second
• Latency — average response time
• Storage — disk usage
• Memory — memory usage
• CPU — CPU usage

6. Cluster Diagnostics

• Check health — check cluster health
• Check config — verify the configuration
• Check logs — find errors in the logs
• Check version — version compatibility

7. Cluster Configuration

Manage the cluster configuration:

• Edit configuration — modify PD, TiDB, and TiKV configs
• View configuration — view the current config
• Restart component — restart a specific node

Advanced Features

Continuous Profiling

Collects profiling data in real time:

-- Enable profiling
SET GLOBAL tidb_enable_pprof = 1;

Top SQL

Displays SQL statements with the highest resource consumption:

• CPU usage per SQL
• Memory usage per SQL
• I/O usage per SQL

Access Logs

View access logs for debugging:

• Client connection logs
• Slow query logs
• Error logs

Monitoring with Grafana

TiDB also provides monitoring via Prometheus + Grafana:

• Grafana: http://:3000
• Prometheus: http://:9090

Main Dashboard Panels

In the next chapter, we will look at monitoring in more detail.