Monitoring vDB with vMonitor Platform

Monitoring is an essential requirement for any system. On VNG Cloud, you can monitor vDB by using the vMonitor Platform service.

vMonitor Platform is a Monitoring as a Service (MaaS) that collects health metrics and logs of resources on VNG Cloud (or outside VNG Cloud). Additionally, you can create visual dashboards, set alarms, and send notifications when incidents occur. For more details about this service, refer to: vMonitor Platform.

To get started, access the vMonitor Platform, go to the Infrastructure List, and select the vDB tab at the link: vMonitor Infrastructure List > vDB. You can refer to the description at: Working with vDB - Metric.

Currently, vMonitor Platform allows you to view the metrics of all vDB Instances with a default dashboard (automatically generated by the system with a limited set of metrics) and a retention period of 1 day, completely free. To create dashboards with an unlimited number of metrics, view metrics with longer retention periods, and create alarms for resource usage (e.g., CPU, memory, disk, connections, buffer, etc.) when they reach dangerous thresholds, you need to Enable detailed Monitoring. This requires you to register for a quota with vMonitor Platform. For more information, refer to: Working with Metric Quota.

After enabling detailed monitoring, you can clone the default dashboard and add other widgets as needed or create new dashboards.

At this point, you can select metrics for vDB in the format: vdb.<engine_name>.<metric_name>. For example: vdb.cpu.percent, vdb.mysql.net.threads_connected, vdb.postgresql.rows_inserted, etc.

The vDB metrics support the following dimensions to filter metrics by specific vDB Instance details such as ID, name, engine type, and version:

• database_id: ID of the vDB Instance.

• database_name: Name of the vDB Instance.

• engine: Type of engine (e.g., MySQL, MariaDB, PostgreSQL, Redis).

• version: Version of the engine.

• zone: Zone of the resource (e.g., HCM-03).

You can also create alarms based on these metrics. Set Alarms for Metrics

For critical vDB instances, VNG Cloud recommends creating at least the following alarms:

• CPU usage: Percentage of CPU usage for the vDB Instance. Use the metric vdb.cpu.percent, with warning thresholds like >80%, >90%, >100%.

• Memory usable: Percentage of available memory for the vDB Instance. Use the metric vdb.mem.usable_perc, with warning thresholds like <20%, <10%.

• Disk usage: Percentage of disk usage for the vDB Instance. Use the metric vdb.disk.space_used_perc, with warning thresholds like >80%, >90%, >100%.