Help Center/ GeminiDB/ GeminiDB DynamoDB-Compatible API/ Best Practices/ Suggestions on Alarm Rules for GeminiDB DynamoDB-Compatible API Metrics

Updated on 2025-07-03 GMT+08:00

View PDF

Suggestions on Alarm Rules for GeminiDB DynamoDB-Compatible API Metrics

After setting alarm rules on the Cloud Eye console, for example, specifying monitored objects and notification policies, you can stay on top of your instance status. For details, see Configuring Alarm Rules.

This section describes recommended alarm rules of GeminiDB DynamoDB-Compatible API.

**Table 1** Suggestions on alarm rules for GeminiDB DynamoDB-Compatible API metrics
Metric ID	Metric Name	Description	Dimension	Threshold (Raw Value) in Best Practices	Alarm Severity in Best Practices	Alarm Handling Suggestion
nosql005_disk_usage	Storage Space Usage	Storage usage Unit: %	Instance	> 80% for 3 consecutive periods	Major	Evaluate how much storage needs to be added based on data growth. For details, see Manually Scaling Up Storage Space. Enable autoscaling. For details, see Automatically Scaling Up Storage Space.
cassandra001_cpu_usage	CPU Usage	CPU usage of an instance Unit: %	Node	> 80% for 3 consecutive periods	Major	Upgrade CPU specifications. For details, see Changing vCPUs and Memory.
cassandra002_mem_usage	Memory Usage	Memory usage of an instance Unit: %	Node	> 80% for 3 consecutive periods	Major	Upgrade memory specifications. For details, see Changing vCPUs and Memory.
cassandra015_read_latency	Average Read Latency	Average latency of database reads Unit: ms	Node	> 900 ms for 3 consecutive periods	Major	Check whether the service traffic increases sharply and whether the database is normal. For details, see Viewing Metrics.
cassandra016_write_latency	Average Write Latency	Average latency of database writes Unit: ms	Node	> 900 ms for 3 consecutive periods	Major	Check whether the service traffic increases sharply and whether the database is normal. For details, see Viewing Metrics.
cassandra037_pending_write	Suspended Write Tasks	Write tasks waiting in the queue Unit: count	Node	> 3,000 for 3 consecutive periods	Major	Check whether the service traffic increases sharply and whether the database is normal. For details, see Viewing Metrics.
cassandra038_pending_read	Suspended Read Tasks	Read tasks in the waiting status Unit: count	Node	> 3,000 for 3 consecutive periods	Major	Check whether the service traffic increases sharply and whether the database is normal. For details, see Viewing Metrics.