This topic describes the metrics for the graphs in a Graph Compute instance.
Metrics page
After you create graphs in a Graph Compute instance, you can go to the Metrics page of the instance details page. On the Metrics page, you can select a graph and specify a time range to view the metrics of the graph and the metrics of the tables in the graph.
Metrics of a graph
The monitored object is the Proxy. The following metrics are displayed for a graph:
Online service queries per second (QPS)
Online service latency
Metrics of the tables in a graph
The monitored objects are all the tables in a graph. The following metrics are displayed:
QPS: the QPS of the Searcher.
Server RT: the server response time (RT) of the Searcher.
Total update QPS
Update QPS with failures
Update time gap: the period from the time when you send an update message to the time when the update message takes effect.
Returned records
Disk usage: the disk space used by different online clusters, which includes data that is mutually backed up.
Memory usage
Real-time memory status