PolarDB-X is a Database Platform as a Service (dbPaaS) built on Kubernetes, available on Alibaba Cloud and hybrid cloud environments. This document covers deployment modes, automated O&M, resource monitoring, SQL diagnostics, backup and restore, and data integration.
Deployment modes
PolarDB-X supports two deployment modes:
| Mode | Description |
|---|---|
| Alibaba Cloud (public cloud) | Fully managed service. Use the unified console or OpenAPI to manage instances. No infrastructure setup required. |
| Hybrid cloud | Deploy on Apsara Stack Enterprise or DBStack for on-premises or industry-specific clouds. DBStack also supports a lightweight mode. Production deployments on DBStack require at least three servers. |
Automated O&M
The PolarDB-X console covers the full O&M lifecycle:
Create and release instances
Back up and restore data
Manage configurations, users, permissions, and parameters
Set up monitoring and alert rules
Analyze and optimize SQL queries
Resource monitoring
The PolarDB-X console displays real-time monitoring data for compute nodes and storage nodes. Create alert rules based on any of these metrics to get notified before issues affect your workload.
Compute node metrics
| Metric | Description |
|---|---|
| CPU utilization | Percentage of CPU capacity in use |
| Memory usage | Percentage of memory in use |
| Network traffic | Inbound and outbound throughput |
| QPS | Queries per second |
| Query latency | Average time to execute a query |
| Frontend connections | Number of active client connections |
| Active threads | Number of threads currently executing |
Storage node metrics
| Metric | Description |
|---|---|
| CPU utilization | Percentage of CPU capacity in use |
| Memory usage | Percentage of memory in use |
| Disk space usage | Storage capacity consumed |
| Network traffic | Inbound and outbound throughput |
| Connections | Total number of connections |
| Active threads | Number of threads currently executing |
| IOPS | Input/output operations per second |
| QPS | Queries per second |
| TPS | Transactions per second |
| Buffer pool usage | InnoDB buffer pool utilization |
| Temporary tables | Number of temporary tables created |
SQL diagnostics and optimization
PolarDB-X provides three tools for diagnosing SQL performance issues. Choose the one that matches your scenario:
| Tool | Use when |
|---|---|
| SQL audit and analysis | You need real-time visibility into SQL query logs and want multi-channel alerts for business exceptions |
| 10-second SQL analysis | You suspect a sudden performance spike and want a snapshot of which queries are running |
| Slow query logs | You want to identify and optimize queries that consistently degrade performance |
SQL audit and analysis
Enable SQL audit and analysis in the console to query logs in real time and analyze them through a visualized interface. Configure custom alert rules to receive notifications through multiple channels when key business metrics exceed thresholds.
View and terminate active sessions from the console. In the Session Statistics section, browse session data aggregated by client or by namespace.
10-second SQL analysis
When enabled, PolarDB-X collects session data once per second for 10 seconds, then summarizes the results. The analysis shows which SQL statements run most frequently and flags slow queries within that window.
Slow query logs
The slow query log UI shows:
Query templates and how often each runs
Execution time per query
Rows returned per query
For each slow query, PolarDB-X analyzes the execution plan and suggests optimizations—such as creating a local index or a global index—to reduce latency.
SQL throttling
If a small number of queries consume a disproportionate share of resources, use SQL throttling to limit their concurrency. Set a maximum number of concurrent queries and a maximum request queue length directly in the console.
Backup and restore
PolarDB-X provides three mechanisms for recovering from data loss or accidental changes:
| Mechanism | Use when |
|---|---|
| Consistent backup and restore | You need to recover an instance to a specific point in time—for disaster recovery, data audits, or rolling back write errors |
| Table recycle bin | A table was accidentally dropped or truncated and needs to be recovered quickly |
| SQL flashback | A specific DML statement (such as DELETE) ran accidentally and you need to reverse it |
Consistent backup and restore
Consistent backup supports point-in-time recovery (PITR) with second-level precision. It combines periodic full snapshot backups with incremental binary log backups to minimize recovery time and data loss.
Table recycle bin
When the table recycle bin is enabled, dropped or truncated tables move to the recycle bin instead of being permanently deleted. Recover a table within two hours of deletion by going to the Table Recycle Bin page in the PolarDB-X console.
SQL flashback
SQL flashback generates a compensating SQL statement that reverses an accidental DML operation. For example, if you accidentally ran a DELETE statement, SQL flashback generates the corresponding INSERT statement to restore the deleted rows.
Data integration
Change Data Capture (CDC)
PolarDB-X uses a Change Data Capture (CDC) component to record data changes in a format compatible with MySQL binary logs. CDC abstracts distributed database internals—such as instance scaling, distributed transactions, and global indexes—so a PolarDB-X cluster behaves like a standard MySQL instance for downstream consumers.
Use open-source tools like Canal to stream data from PolarDB-X to other storage systems.
MySQL ecosystem compatibility
PolarDB-X is compatible with MySQL communication protocols and SQL syntax. Tools that work with MySQL work with PolarDB-X, including:
mysqldumpfor logical backups and migrationsLOAD DATA INFILEfor bulk data imports
PolarDB-X also integrates with the following Alibaba Cloud services:
Data Management (DMS)
Data Transmission Service (DTS)
Database Backup (DBS)
Database Autonomy Service (DAS)
DataWorks