Viewing Table Metrics in the Control System

Explains how to view primary and secondary table index metrics using the Control System.

A subset of the following primary table and secondary index metrics are available as charts and lists in the Control System. For information on how to:
Chart/Column Name Metric Description
Table Bytes Read Per Node Throughput - bytes read The number of bytes read from the primary table per node for all RPC types.
Table and Index Bytes Read The number of bytes read across a primary table and its secondary indexes for all RPC types.
Table Bytes Written Per Node Throughput - bytes written The number of bytes written to the primary table per node for all RPC types.
Table and Index Bytes Written The number of bytes written across a primary table and its secondary indexes for all RPC types.
Table Rows Read Per Node Throughput - rowCount read The number of rows read from the primary table per node for all RPC types.
Table and Index Rows Read The number of rows read across a primary table and its secondary indexes for all RPC types.

This is displayed in the default list view for a node.

Table Rows Written Per Node Throughput - rowCount written The number of rows written to the primary table per node for all RPC types.
Table and Index Rows Written The number of rows written across a primary table and its secondary indexes for all RPC types.
All Tables Written Rows Throughput Number of rows written by RPC operation type.
Table Rows Responded Per Node Throughput - rowCount returned The number of rows returned from the primary table per node for all RPC types.
Table and Index Rows Responded The number of rows returned across a primary table and its secondary indexes for all RPC types.

This is displayed in the default list view for a node.

Scan Throughput Compares the scan throughput for rows read versus rows returned.

This is displayed in the default chart view for a node and for a table.

All Tables Throughput Throughput - rpcCount/second Number of RPC operations by type for all tables in the cluster.

This is displayed in the default chart view for all the tables.

Throughput by Rpc Type The combined RPC load for the primary table and its indexes.
All Tables RPC Byte Throughput The number of bytes processed by RPC operation type.
All Tables Read Rows Throughput Number of rows processed by RPC operation type.
All Tables Returned Rows Throughput Number of rows returned by RPC operation type.
Put and Append Operation Throughput Number of put and append RPC operations for the table, including its indexes.
Table Check and Put Ops Per Node The number of check and put operations completed for a primary table and for a node.
Table Update and Get Ops Per Node The number of update and get operations completed for a primary table and for a node.
Table Get and Index Scans The number of get and index scans completed for a primary table and for a node.
Table Write and Index Maintenance Activity The number of table writes (puts, appends, increments, check and puts, update and gets) that require puts to the index.
Table Get Ops Per Node The number of get operations completed for a primary table and for a node.

This is displayed in the default list view for a node and for a table.

Table Get Throughput Per Node The number of get operations completed per second for a table, excluding its secondary indexes, per node.

This is displayed in the default chart view for a node and for a table.

Table Increment Ops Per Node The number of increment operations completed for a primary table and for a node.
Table Put Ops Per Node The number of put operations completed for a primary table and for a node.

This is displayed in the default list view for a node and for a table.

Table Scan Ops Per Node The number of scan operations completed for a primary table and for a node.
Table Append Ops Per Node The number of append operations completed for a primary table and for a node.
Table Write Throughput Per Node The number of put and append operations per second completed for a table, excluding its secondary indexes, per node.

This is displayed in the default chart view for a node and for a table.

Table and Index Put Ops Per Node The number of put operations completed per secondary index of the primary table.
Table and Index Put Ops The number of put operations completed for the primary table and all its secondary indexes.
Table and Index Scan Ops The number of scan operations completed for the primary table and all its secondary indexes.
Table and Index Scan Ops Per node The number of scan operations completed per secondary index of the primary table.
Table and Index Scan Latency Latency The 99th percentile latency of all scan operations across the primary table and its secondary indexes. A bad ratio between rows read and responded with high scan latency may indicate a poorly configured index.
Table and Index Scan Latency Per node The 99th percentile latency of scan operations completed across the primary table and secondary index per node. Large scans may hit the disks and result in poor performance, or a degrading disk may spike the latency.
Table Append Latency Per Node The 99th percentile latency of append operations on the primary table per node.
Table Increment Latency Per Node The 99th percentile latency of increment operations on the primary table per node.
Table Put Latency Per Node The 99th percentile latency of put operations on the primary table per node.

This is displayed in the default list view for a node and for a table.

Table Get Latency Per Node The 99th percentile latency of get operations on the primary table per node.

This is displayed in the default list view for a node and for a table.

Table Scan Latency Per Node The 99th percentile latency of scan operations on the primary table per node.
Table Get Latency Percentiles* The get operation latency by percentile for the primary table and its secondary indexes.
Primary Table Put & Append Latency Percentiles* The pure write operation latency by percentile for the primary table and its secondary indexes.
Table Write Throughput Latency Percentiles The 99th percentile latency of put operations on the primary table per node.

This is displayed in the default chart view for a node and for a table.

Table Get Throughput Latency Percentiles The 99th percentile latency of get operations on the primary table per node.

This is displayed in the default chart view for a node and for a table.

Table Check and Put Latency Per Node The 99th percentile latency of check and put operations on the primary table per node.
Table Update and Get Latency Per Node The 99th percentile latency of update and get operations on the primary table per node.
Index Put Latency The 99th percentile latency of put operations per secondary index of the primary table.
Index Put Latency Per Node The 99th percentile latency of put operations per secondary index of the primary table per node.
Index Put Latency Percentiles The put operation latency by percentile for all indexes.
Index Scan Latency The 99th percentile latency of scan operations per secondary index of the primary table.
Index Scan Latency Per Node The 99th percentile latency of scan operations per secondary index of the primary table per node.
Index Scan Latency Percentiles The scan operation latency by percentile for all indexes.
All Tables Replication Sent Bytes Replication - rows/bytes sent The number of bytes of replication data sent.
All Tables Replication Pending Bytes Replication - rows/bytes pending The number of bytes of replication data not yet sent.
All Tables Replication Activity Replication - rows/bytes sent/pending Number of bytes of replication data sent vs not yet sent.

This is displayed in the default chart view for all the tables.

Index Throughput by RPC Type Index - Throughput The combined RPC load for the primary table and its secondary indexes.
Index Put Ops The number of put operations completed per secondary index of the primary table.
Index Put Ops Per Node The number of put operations completed per secondary index of the primary table per node.
Index Scan Ops The number of scan operations completed per secondary index of the primary table.
Index Scan Ops Per Node The number of scan operations completed per secondary index of the primary table per node.
Index Bytes Read Index - rows/bytes read The number of bytes read per secondary index of the primary table for all RPC types.
Index Bytes Read Per Node The number of bytes read per secondary index of the primary table per node for all RPC types.
Index Bytes Written Per Node The number of bytes written per secondary index of the primary table per node for all RPC types.
Index Rows Read The number of rows read per secondary index of the primary table for all RPC types.
Index Rows Read Per Node The number of rows read per secondary index of the primary table per node for all RPC types.
Index Rows Responded Index - rows/bytes returned The number of rows returned per secondary index of the primary table for all RPC types.
Index Rows Responded Per Node The number of rows returned per secondary index of the primary table per node for all RPC types.
Index Scan Read vs Returned Rows Index - rows/bytes read The number of secondary index rows that were read versus returned.
Index Bytes Written Index - rows/bytes write The number of bytes written per secondary index of the primary table for all RPC types.
Index Rows Written The number of rows written per secondary index of the primary table for all RPC types.
Index Rows Written Per Node The number of rows written per secondary index of the primary table per node for all RPC types.
All Tables Index Sent Bytes Index - rows/bytes sent The number of bytes sent for secondary index updates.
All Tables Index Pending Bytes Index - rows/bytes pending The number of bytes of secondary index data remaining to be sent.
All Index Maintenance Activity Number of bytes of index data sent vs not yet sent.

This is displayed in the default chart view for all the tables.

All Tables CDC Sent Bytes CDC - rows/bytes sent The number of bytes of CDC data sent.
All Tables CDC Pending Bytes CDC - rows/bytes pending The number of bytes of CDC data per node not yet sent.
All Tables CDC Propagation Activity The number of bytes of CDC data sent vs not yet sent.

This is displayed in the default chart view for all the tables.

All Streams Producer Ops Streams Throughput, RPCs The number of Streams producer RPCs.
All Streams Consumer Ops The number of Streams consumer RPCs.
All Streams Producer Messages Streams Throughput, messages The number of Streams messages produced.
All Streams Consumer Messages The number of Streams messages read by consumers.
Table Value Cache All Lookups Value Cache Lookups All operations for a primary table and for a node that performed a cache lookup.
Table Value Cache Lookups The number of get operations for a primary table and for a node that performed a cache lookup.
Table Value Cache Lookups Per Index The number of get operations for a primary table and for a node that performed a cache lookup.
Table and Index Value Cache All Lookups Value Cache Hits All operations across the primary table and its secondary indexes that performed a cache lookup.
Table and Index Value Cache Lookups Per Index The number of get operations across the primary table and its secondary indexes that performed a cache lookup per secondary index.
Table and Index Value Cache Lookups The number of get operations across the primary table and its secondary indexes that performed a cache lookup.
Table Value Cache All Hits All operations for a primary table and for a node that resulted in a cache hit.
Table Value Cache Hits The number of get operations for a primary table and for a node that resulted in a cache hit.
Table Value Cache Hits Per Index The number of get operations for a primary table and for a node that resulted in a cache hit.
Table and Index Value Cache All Hits All operations across the primary table and its secondary indexes that resulted in a cache hit.
Table and Index Value Cache Hits The number of get operations across the primary table and its secondary indexes that resulted in a cache hit.
Table and Index Value Cache Hits Per Index The number of get operations across the primary table and its secondary indexes that resulted in a cache hit per secondary index.
Value Cache Utilization Value Cache Compares the relative distribution of get operations that either hit the cache or require a lookup.

This is displayed in the default chart view for a node and for a table.

All Tables Flushes Bucket Flushes The number of table flushes that were manually and automatically triggered. Table flushes reorganize data from bucket files (unsorted data) to spill files (sorted data) when the bucket size exceeds a threshold.

This is displayed in the default chart view for all the tables.

All Tables Flushes The number of total table flushes that were manually versus automatically triggered.
All Tables Force Flushes Bucket Force Flushes The number of table flushes that were not automatically triggered.
All Tables Compactions Compaction Number of table compactions. Compactions combine multiple MapR Database data files containing sorted data (known as spills) into a single spill file.

This is displayed in the default chart view for all the tables.

All Tables Full Compactions Number of full compactions. Compactions combine multiple MapR Database data files containing sorted data (known as spills) into a single spill file. Full compactions improve read performance because after compaction, MapR Database needs to read only the single resulting sorted spill file. But they incur I/O costs because the compaction must read, sort, and rewrite all data in the spill files.
All Tables Mini Compactions Number of partial compactions. Compactions combine multiple MapR Database data files containing sorted data (known as spills) into a single spill file. After a mini compaction, MapR Database needs to read only two spill files.
All Tables TTL Compactions Number of compactions that result in reclamation of disk space after removal of stale data. You can control the frequency of TTL compactions by configuring the TTL for a table's column families.
All Tables Free Index Memory Memindex Usage The number of available MB in the in-memory bucket file cache.

* Percentiles are estimated by linearly interpolating between fixed buckets sizes.