org.apache.cassandra.metrics:type=Table keyspace=<Keyspace> scope=<Table> name=<MetricName>
Metric Name Tye Description
--------------------------------------------------------------------------
ReadLatency Latency Local read latency for this table.
RangeLatency Latency Local range scan latency for this table.
WriteLatency Latency Local write latency for this table.
CoordinatorReadLatency Timer Coordinator read latency for this table.
CoordinatorWriteLatency Timer Coordinator write latency for this table.
CoordinatorScanLatency Timer Coordinator range scan latency for this table.
1条答案
按热度按时间ilmyapht1#
tablestats
将为您提供可能足够的请求总数。还可以查看那里的平均本地读/写延迟以查找异常值。费率在jmx中公开,您可以从表度量中获取:http://cassandra.apache.org/doc/latest/operating/metrics.html#table-指标其中每个都有1、5和15分钟速率属性。
使用瑞士 java 刀:
(注意:以与您的cassandra示例运行时相同的用户身份运行它
sudo -u
或者它可能没有附加到jvm的权限)如果它目前运行高负荷,你可以使用
toppartitions
或者profileload
. 在某些版本中,toppartitions要求您将表交给它。