Usage Guide
Assume that the metric collection system is running properly and the configuration file directory confpath has been initialized. You can run the following command to implement this feature:
Enable only the anomaly detection function:
gs_dbmind service start --conf confpath --only-run anomaly_detection
View a metric on all nodes from timestamps1 to timestamps2:
gs_dbmind component anomaly_detection --conf confpath --action overview --metric metric_name --start-time timestamps1 --end-time timestamps2
View a metric on a specific node from timestamps1 to timestamps2:
gs_dbmind component anomaly_detection --conf confpath --action overview --metric metric_name --start-time timestamps1 --end-time timestamps2 --host ip_address --anomaly anomaly_type
View a metric on all nodes from timestamps1 to timestamps2 in a specific anomaly detection mode:
gs_dbmind component anomaly_detection --conf confpath --action overview --metric metric_name --start-time timestamps1 --end-time timestamps2 --anomaly anomaly_type
View a metric on a specific node from timestamps1 to timestamps2 in a specific anomaly detection mode:
gs_dbmind component anomaly_detection --conf confpath --action overview --metric metric_name --start-time timestamps1 --end-time timestamps2 --host ip_address --anomaly anomaly_type
Visualize a metric on all nodes from timestamps1 to timestamps2 in a specific anomaly detection mode:
gs_dbmind component anomaly_detection --conf confpath --action plot --metric metric_name --start-time timestamps1 --end-time timestamps2 --host ip_address --anomaly anomaly_type
Stop the started service:
gs_dbmind service stop --conf confpath
NOTE: When setting anomaly detection parameters, ensure that start-time is at least 30 seconds earlier than end-time.