Alarm Upgrade

Last updated: 2021-06-17 14:42:27

Background

TencentDB for MariaDB upgraded the monitoring items for server and component services on April 1, 2021 by replacing the legacy alarm policy type and modifying hundreds of monitoring and alarm metrics. You can configure alarm policies of the MariaDB type in the Cloud Monitor console.

The original MariaDB (legacy) policy type will no longer be maintained in the future, and the backend will gradually migrate all the MariaDB policies you previously configured to the new alarm policy type. We recommend you migrate the alarms on your own, so that you can try out the new alarm feature earlier. Subsequently, please configure new alarm policies under the new alarm policy type.

Comparison of the legacy and new alarm policy types

Policy Type Metric Coverage Support and Maintenance
MariaDB (legacy) 22 metrics You can continue to use this policy type, but it will no longer be maintained or updated, and your instances will be automatically migrated to the new policy type gradually.
MariaDB 37 metrics This policy type was released on April 1, 2021 with ongoing maintenance available.
Note:

Alarm Policy Migration

After the MariaDB (legacy) policy type is deactivated, the system will automatically migrate previously configured alarm policies to the new MariaDB policy type on the backend.

Note:

Alarms may not be automatically migrated to the new alarm policy type by the system for certain instances or users. If this is the case for you, we will notify you through Message Center, email, or SMS. Then, please follow the manual migration steps below to manually migrate the alarms.

Manual migration steps

  1. Sort out exiting alarm metrics and policies.
    1. Log in to the Cloud Monitor console, select Alarm Configuration > Alarm Policy on the left sidebar, and click Advanced Filter.
    2. On the pop-up page, select the alarm policy type corresponding to MariaDB (legacy) in Policy Type, query alarm policies in this category, and download the previously configured alarm policies of the original MariaDB (legacy) policy type.
  2. Configure new alarm policies.
    1. On the Alarm Policy page, click Create.
    2. On the Create Alarm Policy page, select MariaDB for Policy Type and configure alarms according to the policies downloaded in step 1. For the configuration method, please see Creating Alarm Policy.
  3. Verify whether the MariaDB alarm policies are enabled and can successfully trigger alarms.
    Set a minimum trigger threshold in Metric alarm on the Create Alarm Policy page, choose to set a recipient or recipient group, and select the notification channel (email or SMS) to test a policy. For example, you can configure an alarm policy for the CPU utilization metric that triggers an alarm once per minute when the threshold is greater than or equal to 1% for one statistical period of one minute.
  4. After the new policy type is verified, delete the alarm policies previously configured under the original MariaDB (legacy) policy type.
    On the Alarm Policy page, filter alarm policies by the "MariaDB (legacy)" policy type and delete the filtered policies according to the policy list downloaded in step 1.
    If you encounter any issues during the migration, please submit a ticket for assistance.

Comparison Table of New and Legacy Metrics

Original Policy TypeMetric/Event AlarmOriginal Metric/Event Alarm NameNew Policy TypeNew Metric/Event Alarm Name
TencentDB for MariaDB (legacy) Metric alarmCPU utilization MariaDB - instance summaryCPU utilization
Metric alarmPrimary/Replica switch MariaDB - instance summaryNumber of primary/replica switches
Metric alarm Available disk space MariaDB - instance summary Available data disk space
Metric alarm Log disk space MariaDB - instance summary Minimum remaining binlog disk space
Metric alarm Query count (SELECT) MariaDB - instance summary Aggregated number of SELECT requests
Metric alarm Slow query count MariaDB - instance summary Slow query count
Metric alarm Update count (UPDATE) MariaDB - instance summary Number of UPDATE requests
Metric alarm Insertion count (INSERT) MariaDB - instance summary Number of INSERT requests
Metric alarm Deletion count (DELETE) MariaDB - instance summary Number of DELETE requests
Metric alarm Available memory size MariaDB - instance summary Available cache space
Metric alarm Disk IOPS MariaDB - instance summary IO utilization
Metric alarm Number of active connections MariaDB - instance summary Aggregated number of active threads
Metric alarm Connection utilization MariaDB - instance summary Maximum database connection utilization
Metric alarm Disk utilization MariaDB - instance summary Data disk space utilization
Metric alarm Replacement count MariaDB - instance summary Number of REPLACE_SELECT requests
Metric alarm Number of InnoDB disk reads MariaDB - instance summary Aggregated number of InnoDB disk reads
Metric alarm Number of InnoDB buffer pool reads MariaDB - instance summary Aggregated number of InnoDB buffer pool reads
Metric alarm Number of InnoDB buffer pool read-aheads MariaDB - instance summary Aggregated number of InnoDB buffer pool read-aheads
Metric alarm Number of deleted InnoDB rows MariaDB - instance summary Number of deleted InnoDB rows
Metric alarm Number of inserted InnoDB rows MariaDB - instance summary Number of inserted InnoDB rows
Metric alarm Number of updated InnoDB rows MariaDB - instance summary Number of updated InnoDB rows
Metric alarm Used log disk space MariaDB - instance summary Used binlog disk space

New Metric Description

Policy TypeMetric/Event AlarmMetric/Event Alarm Name
TencentDB for MariaDB Metric alarm Aggregated number of read InnoDB rows
Metric alarm Minimum replica delay
Metric alarm Cache hit rate
Metric alarm Number of REPLACE requests
Metric alarm Number of requests lasting for less than 5 ms
Metric alarm Number of requests lasting for 5–20 ms
Metric alarm Number of requests lasting for 21–30 ms
Metric alarm Number of requests lasting for more than 30 ms
Metric alarm Total number of SQL requests
Metric alarm Number of SQL errors
Metric alarmNumber of SQL successes
Metric alarm Total number of connections to clients
Metric alarm Aggregated number of requests to primary/replica nodes
Metric alarm Aggregated number of active connections
Metric alarm Aggregated maximum number of connections