Configure DAS Alert Rules for Threshold & Anomaly Monitoring - Database Autonomy Service

Database Autonomy Service (DAS) lets you create custom alert rules to monitor your database instances. When an instance triggers a rule, DAS automatically sends notifications through your configured channels.

How it works

Alert rules are the building blocks of the DAS alerting system:

Alert rule — defines what to monitor (a metric threshold or an autonomy event) and how to send notifications
Alert template — a collection of alert rules applied to one or more database instances
Alert — the notification sent when a rule triggers

Create alert rules first, then add them to an alert template and apply the template to your database instances. For instructions, see Configure alerts.

Important

The alert types in DAS have been updated. Autonomy Events is now available, and Event Alerting is no longer supported. Existing rules are retained. When creating new rules, select either Threshold Alerting or Autonomy Events.

Prerequisites

Before you begin, ensure that you have:

A DAS-managed database instance (ApsaraDB or a directly connected self-managed database)
Access to the DAS console

Choose an alert type

Select an alert type before creating a rule:

Alert type	Use when
Threshold Alerting	You want to monitor a specific metric (such as CPU utilization or disk usage) and receive alerts when it exceeds a defined value. Best for steady-state monitoring with predictable thresholds.
Autonomy Events	You want to be notified when DAS takes or detects an autonomous action, such as auto scaling, SQL optimization, or a throttling event. Best for tracking DAS-managed interventions.

Create an alert rule

Log on to the DAS console.
In the left navigation pane, click Tools > Alert Service > Alert Rules.
On the Alert Rules page, click Add Rule in the upper-right corner.
In the Add Rule dialog box, configure the following parameters and click OK.

Parameter	Required	Description
Rule name	Yes	A name for the alert rule.
Type	Yes	The alert type. Select Threshold Alerting to trigger alerts when a metric exceeds a specified threshold. Select Autonomy Events to trigger alerts when a specified autonomy event occurs.
Description	If Type is Threshold Alerting	The conditions that trigger the alert. Detection granularity is 1 minute.
Event type	If Type is Autonomy Events	The autonomy event type that triggers the alert. Supported types: Auto Scaling Event, Throttling Event, SQL Optimization, Storage Optimization, and Metric Exception.
Alert interval	Yes	How often alert notifications are sent after a rule triggers.
Effective time	Yes	The time window during which the alert rule is active.
Notification level	Yes	The severity level for the alert notification.
Notification method	Yes	How notifications are delivered: text message, webhook, or email. Webhooks are supported for DingTalk, WeCom, Lark, and Application Real-Time Monitoring Service (ARMS).

After creating the rule, add it to an alert template and apply the template to your database instances. For more information, see Configure alerts.

Manage alert rules

Log on to the DAS console.
In the left navigation pane, click Tools > Alert Service > Alert Rules.
On the Alert Rules page, use the Actions column to manage existing rules.

Action	Description
Search	View the details and alert history of the rule.
Edit	Modify the rule configuration. For parameter details, see Create an alert rule.
Disable	Temporarily disable the rule.
Delete	Permanently delete the rule.

Important

Modifying, disabling, or deleting an alert rule affects all alert templates that contain it. Proceed with caution.

Alert metrics

Important

These alert metrics apply only to Alibaba Cloud database instances.

MySQL

Alert metric	Unit	Description
MySQL disk usage	%	Disk usage.
MySQL CPU utilization	%	CPU utilization of the MySQL service process. The maximum value is 100% for ApsaraDB instances.
MySQL IOPS utilization	%	IOPS utilization.
MySQL memory usage	%	Memory usage of the MySQL instance as a percentage of total operating system memory.
MySQL active sessions	Count	Number of current active sessions.
MySQL transactions per second	Per second	Average number of transactions per second.
MySQL requests per second	Per second	Average number of requests per second.
MySQL DELETE statements per second	Per second	Average number of DELETE statements executed per second.
MySQL INSERT statements per second	Per second	Average number of INSERT statements executed per second.
MySQL INSERT_SELECT statements per second	Per second	Average number of INSERT_SELECT statements executed per second.
MySQL REPLACE statements per second	Per second	Average number of REPLACE statements executed per second.
MySQL REPLACE_SELECT statements per second	Per second	Average number of REPLACE_SELECT statements executed per second.
MySQL SELECT statements per second	Per second	Average number of SELECT statements executed per second.
MySQL UPDATE statements per second	Per second	Average number of UPDATE statements executed per second.
MySQL full table scans per second	Per second	Number of full table scan queries per second.
MySQL tables opened	Count	Number of opened tables.
MySQL files opened	Count	Number of opened files.
MySQL rows sorted per second	Per second	Number of rows sorted per second.
MySQL secondary node replication delay	Second	Replication delay of the secondary node.
MySQL InnoDB average rows deleted per second	Per second	Average number of rows deleted from InnoDB per second.
MySQL InnoDB average rows inserted per second	Per second	Average number of rows inserted into InnoDB per second.
MySQL InnoDB average rows read per second	Per second	Average number of rows read from InnoDB per second.
MySQL InnoDB average rows updated per second	Per second	Average number of rows updated in InnoDB per second.
MySQL InnoDB pages read from Buffer Pool per second	Per second	Average number of pages read from the InnoDB Buffer Pool per second (logical reads).
MySQL InnoDB pages written to Buffer Pool per second	Per second	Average number of pages written to the InnoDB Buffer Pool per second.
MySQL InnoDB Buffer Pool dirty page ratio	%	Ratio of dirty pages in the InnoDB Buffer Pool. Formula: `Innodb_buffer_pool_pages_dirty / Innodb_buffer_pool_pages_data × 100%`.
MySQL InnoDB Buffer Pool read cache hit ratio	%	Read cache hit ratio of the InnoDB Buffer Pool. Formula: `(Innodb_buffer_pool_read_requests - Innodb_buffer_pool_reads) / Innodb_buffer_pool_read_requests × 100%`.
MySQL InnoDB table average row lock waits	Count	Average number of row lock waits on the InnoDB table.
MySQL InnoDB table maximum row lock wait time	ms	Maximum row lock wait time on the InnoDB table.
MySQL instance total space usage	MB	Total space usage of the MySQL instance.

PolarDB for MySQL

Alert metric	Unit	Description
PolarDB for MySQL disk usage	%	Disk usage.
PolarDB for MySQL service process CPU utilization	%	CPU utilization of the service process. A value of 200% indicates that two CPU cores are in use.
PolarDB for MySQL IOPS utilization	%	IOPS utilization.
PolarDB for MySQL instance memory usage	%	Memory usage of the instance as a percentage of total operating system memory.
PolarDB for MySQL active sessions	Count	Number of current active sessions.
PolarDB for MySQL transactions per second	Per second	Formula: `(Com_commit + Com_rollback) / Uptime`.
PolarDB for MySQL requests per second	Per second	Formula: `Queries / Uptime`.
PolarDB for MySQL DELETE statements per second	Per second	Average number of DELETE statements executed per second.
PolarDB for MySQL INSERT statements per second	Per second	Average number of INSERT statements executed per second.
PolarDB for MySQL INSERT_SELECT statements per second	Per second	Average number of INSERT_SELECT statements executed per second.
PolarDB for MySQL REPLACE statements per second	Per second	Average number of REPLACE statements executed per second.
PolarDB for MySQL REPLACE_SELECT statements per second	Per second	Average number of REPLACE_SELECT statements executed per second.
PolarDB for MySQL SELECT statements per second	Per second	Average number of SELECT statements executed per second.
PolarDB for MySQL UPDATE statements per second	Per second	Average number of UPDATE statements executed per second.
PolarDB for MySQL Multi-DELETE statements per second	Per second	Average number of Multi-DELETE statements executed per second.
PolarDB for MySQL Multi-UPDATE statements per second	Per second	Average number of Multi-UPDATE statements executed per second.
PolarDB for MySQL InnoDB rows deleted per second	Per second	Average number of rows deleted from InnoDB per second.
PolarDB for MySQL InnoDB rows inserted per second	Per second	Average number of rows inserted into InnoDB per second.
PolarDB for MySQL InnoDB rows read per second	Per second	Average number of rows read from InnoDB per second.
PolarDB for MySQL InnoDB rows updated per second	Per second	Average number of rows updated in InnoDB per second.

Redis

Alert metric	Unit	Description
Redis memory utilization	%	Memory utilization.
Redis average response time	μs	Average response time.
Redis process CPU utilization	%	CPU utilization of the Redis process.
Redis network read rate per second	KB/s	Network read rate per second.
Redis network write rate per second	KB/s	Network write rate per second.
Redis network read utilization	%	Network read utilization.
Redis network write utilization	%	Network write utilization.

PostgreSQL

Alert metric	Unit	Description
PostgreSQL process CPU utilization	%	CPU utilization of the PostgreSQL process.
PostgreSQL active sessions	Count	Number of current active connections.
PostgreSQL IOPS usage	%	IOPS usage.
PostgreSQL disk usage	%	Disk usage.

PolarDB for PostgreSQL

Alert metric	Unit	Description
PolarDB for PostgreSQL CPU utilization	%	CPU utilization.
PolarDB for PostgreSQL memory usage	%	Memory usage.
PolarDB for PostgreSQL active sessions	Count	Number of current active connections.
PolarDB for PostgreSQL total IOPS	Per second	Total IOPS.

Historical alert metrics

Note DAS began supporting new alert metrics on March 16, 2023. If you configured alerts before this date, the details are shown below. Historical alert metrics are being phased out and can no longer be configured. Use the new alert metrics instead.

Alert metrics

Event alerting

Alert metric	Number of occurrences
Database Disconnected	1, 3, 5, 7, or 10 consecutive times.
MySQL Replication Interrupted	1, 3, 5, 7, or 10 consecutive times.

Threshold alerting

Important

ApsaraDB instances and directly connected self-managed databases use the centralized-mode DBGateway connection type.

Alert metric	Threshold unit	Centralized-mode DBGateway	Host-mode DBGateway	Number of occurrences
Host CPU utilization	%	Not supported	Supported	1, 3, 5, 7, or 10 consecutive times.
Host LOAD_1min	None	Not supported
Host memory usage	%	Not supported
Host network sent bandwidth usage	%	Not supported
Host network received bandwidth usage	%	Not supported
Host root partition disk usage	%	Not supported
MySQL active sessions	Count	Supported
MySQL Connections	Count	Supported
MySQL replication delay	s	Supported
MySQL cache hit ratio	%	Supported
MySQL slow SQL statements per hour	Count	Supported
MySQL data storage usage	%	Supported
MySQL log storage usage	%	Supported
Redis client connections	Item	Supported
Redis cache hit ratio	%	Supported
Redis memory fragmentation ratio	%	Supported
PostgreSQL IOPS usage	%	Supported
PostgreSQL iNode usage	%	Supported
PostgreSQL space usage	%	Supported
PostgreSQL TPS	Count	Supported
PostgreSQL connection usage	%	Supported
PostgreSQL average active connections per CPU	Count	Supported
PostgreSQL longest bloat duration	Unit	Supported
PostgreSQL CPU utilization	%	Not supported
PostgreSQL memory usage	%	Supported
MongoDB Connections	Count	Supported
MongoDB cache usage	%	Supported
MongoDB current queued operations	Unit	Supported
MongoDB cache dirty data percentage	%	Supported
MongoDB slow SQL statements per hour	Count	Not supported	Not supported	N/A

FAQ

Why am I receiving event alerts I didn't configure?

Your database instance likely has a legacy event subscription, which triggers alerts for all autonomy event types. To stop receiving unwanted alerts, configure the specific autonomy event alerts you need — the system automatically migrates the legacy subscription to the new version. After migration, alerts for unconfigured event types are no longer sent.

If you don't need any autonomy event alerts, first configure an autonomy event alert of any type to overwrite the legacy subscription. Wait 15 minutes, then disassociate the alert template from the database instance.

What's next

Configure alerts — apply alert rules to database instances using alert templates