HBase integration in Pulse is a distributed column-oriented database that stores huge amounts of data that can be accessed in a random manner. It also provides random real-time read/write access to data in the Hadoop File System. Pulse provides you with an overview of each component of your HBase database. The following components can be viewed:
- HBase Cluster details
- HBase Regions
- HBase Tables
- HBase Region Servers
The HBase Dashboard provides you with the basic statistics about an HBase cluster.
The default time range is Last 24 hrs. To view metrics for a custom date range, click the Timeframe icon in the top-right corner of the page and select the required time frame and timezone.
The following table describes the Cluster tile.
|Num Region Servers||The total number of region servers in the cluster|
|Region Servers Dead||The total number of region servers that are dead in the cluster|
|Master Status||The state of the HBase Master in the cluster, whether it is running or stopped|
|Cluster Requests||The total number of requests in the cluster|
|Size||The size of the cluster, in GB|
|Req Rate/Sec||The total number of requests per second|
|Read Count||The total number of reads completed in the cluster|
|Write Count||The total number of writes completed in the cluster|
The following table describes the Region tile.
|Num Regions||The number of regions in the cluster|
|Regions Over Threshold||The number of regions in transition state for longer than 1 minute in the cluster|
|Regions in Transition||The number of regions in transition in the cluster|
|WAL File Count||The number of Write-Ahead-Log files in the cluster|
|WAL File Size||The total files size of Write-Ahead-Log in the cluster|
|Store File Count||The total number of store files in the cluster|
|Store File Size||The total size of store files in the cluster|
|Files Local %||Percentage of files served from the local Data Node for the region server in the cluster|
The following table describes the Region Compaction tile.
|Avg Compaction Time||The average time taken to process compaction operations in the region server|
|Compaction Queue Size||The total number of compaction operations waiting to be processed in the region server|
|Num Compaction Running||The total number of compactions being processed in the region server|
The following table describes the Table tile.
|# Tables||The total number of tables in the cluster|
|Table Size [Min, Max]||The total size of the data for the table|
|Table Request [Min, Max]||The total number of read and write requests per second for the table|
The following table describes the Cache Metrics tile.
|Hit Ratio||The block cache hit ratio (0 to 100%) from Region Server|
|Avg Flush Time||The average latency for Flush Operation for the tables|
|Flush Queue Size||The size of the output file for Flush Operations for the tables|
The following table describes the Regions Stats tile.
|Online||The number of regions online|
|Offline||The number of regions offline |
You cannot read and write on Offline regions.
If you disable any table, all regions in that table will move to offline state.
|Failed||The number of failed regions |
All regions that are not in open state due to region metadata corruption or failure in the region server are considered as failed regions.
|Split||The number of split regions. Regions split when they reach the configured threshold.|
|Others||The number of regions that are not in the above categories|
|Total||The sum of all the metrics in the Region Stats tile|
Charts on the Hbase Dashboard
The following table describes all the charts displayed in the HBase Dashboard.
|Region & Table Distribution||Displays the following charts: |
|Cluster JVM Usage (95th percentile)||Region server Heap plays a critical role in holding Hbase Region’s memstore and high heap utilization, can cause frequent flush which can impact write latencies directly as it involves disk i/o. |
Click Show Individual to view individual JVM Usage metrics in the region server.
Click Show Aggregated to view a consolidated chart of all JVM Usage metrics in the HBase cluster.
|Request Count||Request Count chart is a time series chart that displays the number of requests(reads and writes) sent by all the regions servers at a particular time. |
Click Show Individual to view individual server region request count in the region server.
Click Show Aggregated to view a consolidated chart of all request counts for all region servers in the HBase cluster.
|Region Latency Timeline (95th percentile)||It is a time series chart that displays the time taken by HBase region servers to process your request. Click the down arrow key to select the following request types: |
|Region Load||It is a histogram chart displaying the read and write request count for all region servers in the cluster.|
|Connections||It is a bar chart displaying the number of open connections in a region server.|
|Request Handlers||It is bar chart displaying the number of request handlers in a region server.|
|Compaction Times & Duration||It is a histogram chart displaying the number of compactions run and the duration of the compaction|
|RPC Time & Process Time - 95th percentile||It is a histogram chart displaying the processing time and queue time of a region server.|
Compaction Region Server
The following table describes the Compaction Region Server metrics.
|Region Server||The name or ID of the region server|
|Compacting KVs||The number of key values in compacting state|
|Compacted KVs||The number of key values that have finished compacting|
|Compaction Progress||The compaction progress, expressed as a percentage|