[HBASE-27681] Refactor Table Latency Metrics - ASF JIRA

XML

Word

Printable

JSON

Details

Type: Improvement
Status: Resolved
Priority: Major
Resolution: Fixed
Affects Version/s: None
Fix Version/s: 2.6.0, 3.0.0-alpha-4
Component/s: None
Labels:
None

Release Note:

Hide
Table request metrics (table latency histograms and TableReadQueryMeter/TableWriteQueryMeter) have been reorganized into a new metric bean TableRequests. The new TableRequests bean is published separately for each table on the regionserver, with the bean name including the table name and metric names within the bean simplified. Beans will be cleaned up after tables leave the regionserver.

Users who collect TableLatencies or table query meters will have to update metric collection to work with the new structure. For example:

Old:
{
      "name": "Hadoop:service=HBase,name=RegionServer,sub=TableLatencies",
      "modelerType": "RegionServer,sub=TableLatencies",
      "tag.Context": "regionserver",
      "tag.Hostname": "hostname",
      "Namespace_default_table_usertable_metric_deleteTime_num_ops": 0,
      "Namespace_default_table_usertable_metric_deleteTime_min": 0,
      "Namespace_default_table_usertable_metric_deleteTime_max": 0,
      "Namespace_default_table_usertable_metric_deleteTime_mean": 0,
      ...
}

New:
{
      "name": "Hadoop:service=HBase,name=RegionServer,sub=TableRequests_Namespace_default_table_usertable",
      "modelerType": "RegionServer,sub=TableRequests_Namespace_default_table_usertable",
      "tag.Context": "regionserver",
      "tag.Hostname": "hostname",
      "DeleteTime_num_ops": 0,
      "DeleteTime_min": 0,
      "DeleteTime_max": 0,
      "DeleteTime_mean": 0,
      ...
}

Show
Table request metrics (table latency histograms and TableReadQueryMeter/TableWriteQueryMeter) have been reorganized into a new metric bean TableRequests. The new TableRequests bean is published separately for each table on the regionserver, with the bean name including the table name and metric names within the bean simplified. Beans will be cleaned up after tables leave the regionserver. Users who collect TableLatencies or table query meters will have to update metric collection to work with the new structure. For example: Old: {       "name": "Hadoop:service=HBase,name=RegionServer,sub=TableLatencies",       "modelerType": "RegionServer,sub=TableLatencies",       "tag.Context": "regionserver",       "tag.Hostname": "hostname",       "Namespace_default_table_usertable_metric_deleteTime_num_ops": 0,       "Namespace_default_table_usertable_metric_deleteTime_min": 0,       "Namespace_default_table_usertable_metric_deleteTime_max": 0,       "Namespace_default_table_usertable_metric_deleteTime_mean": 0,       ... } New: {       "name": "Hadoop:service=HBase,name=RegionServer,sub=TableRequests_Namespace_default_table_usertable",       "modelerType": "RegionServer,sub=TableRequests_Namespace_default_table_usertable",       "tag.Context": "regionserver",       "tag.Hostname": "hostname",       "DeleteTime_num_ops": 0,       "DeleteTime_min": 0,       "DeleteTime_max": 0,       "DeleteTime_mean": 0,       ... }

Description

Benefit:

Table Latency Metrics could removed after table has been moved away. Fix ~~HBASE-27617~~
Could remove the hash lookup caused by metrics map from the hot request path.
Reduce the output in jmx. As if we use jmx_exporter to collect metrics to Prometheus, the operation overhead of performing regular matching in metric is relatively high(especially region metrics, which might be next step).
I think that could be the first step after hbase-metrics released. It seems that the roadmap indicates that we should replace directly use hadoop-metrics2 with hbase-metrics.

Influence: