The following is a list of frequently used stats available from

Statistics are available as either counters or gauges. Counters are additive whereas a gauge captures a statistic at a given point of time and can be used to graph movement of that statistic.

Use the REGEXP function to query the following statistics. This example queries transactions per second from the prior 24 hours:

sql> USE clustrix_statd;
sql> SELECT timestamp, value, name 
     FROM clustrix_statd.statd_history 
     NATURAL JOIN clustrix_statd.statd_metadata 
     WHERE name REGEXP 'tps$' 
     AND timestamp > now() - interval 1 day 
     ORDER BY timestamp DESC; 


Metric NameREGEXPDescriptionMeasuresGraph as
clustrix.stats.connectionsconnectionsTotal number of connections to the clusterNumber of connectionscounter


Metric NameREGEXPDescriptionMeasuresGraph as
clustrix.containers.countcontainers.countNumber of containers in use by XpandContainer Countgauge
clustrix.cluster.nodes_in_quorumnodes_in_quorumNumber of nodes currently online and in the clusterNumber of nodes onlinegauge
clustrix.stats.bigc_value.node.nnbigc_value.node.*XID value through which garbage collection will clean upBigC value by nodegauge
clustrix.stats.fc_wait_time.node.nnfc_wait_time.*Total time all queries spent waiting for flow control (inter-node communication)Time waiting on flow controlcounter
clustrix.stats.key_caches_updates.node.nnkey_caches_updates.node.Number of times the PD Cache issued update keys. This statistic is sometimes useful for support.PD Cache keys updatedcounter
clustrix.stats.layer_merge_bytes.node.nnlayer_merge_bytes.node.Size is bytes of layer merge operations by nodeLayer merge bytes by nodecounter
clustrix.stats.layer_merge_rows.node.nnlayer_merge_rows.node.Number of rows merged by layer merge operations by nodeLayer merge rows by nodecounter
clustrix.stats.pdcache_fetches.node.nnpdcache_fetches.node.Number of times the PD Cache fetched new PDsPD Cache fetches by nodecounter
clustrix.stats.pdcache_loaded.node.nnpdcache_loaded.node.Number of times the PD Cache loaded new PDsPD Cache loaded by nodecounter
clustrix.stats.qpc_missesqpc_missesNumber of times a query was missed by the QPCQuery Plan Cache, missescounter
clustrix.stats.qpc_recompilesqpc_recompilesNumber of times the QPC recompiled queriesQuery Plan Cache, recompilescounter
clustrix.stats.queries_not_using_indexesqueries_not_using_indexesNumber of queries on the cluster not using any indexQueries not using indexescounter


walltime_(read|write)_usTotal wall time (elapsed) of read or write queriesWall time of read or write queriescounter


Metric NameREGEXPDescriptionMeasuresGraph as




Metrics for various Rebalancer actionsVarious Rebalancer actions.counter
clustrix.cluster.total_nodescluster.total_nodesThe total number of nodes configured for your cluster (not just in quorum)Total number of nodes gauge


Metric NameREGEXPDescriptionMeasuresGraph as interface)

latency_ms.avg.node.Network latency average per nodeNetwork Latency avg by nodegauge|max|avg)latency_ms....$Network latency min/max/avgNetwork latency min/max/avggauge


Metric NameREGEXPDescriptionMeasuresGraph as


core0.node.*TIL CPU core-0 utilization by nodeTil CPU Core-0 by nodegauge


til_avg$Average core utilization for the cluster, excluding core-0Average Til CPUgauge
clustrix.activity.til_avg.node.nntil_avg.node.*Average TIL core (non core-0) utilization per node in a clusterAverage Til CPU by nodegauge
clustrix.activity.til_(min|max|avg).node.nntil_(min|max|avg).node.TIL CPU core (non core-0) utilization min/max/avgTil CPU min/max/avg by nodegauge


til_max$Max core utilization for the cluster, excluding core-0Max Til CPUgauge
clustrix.activity.til_max.node.nntil_max.node.*Average TIL core (non core-0) utilization per node in a clusterMax Til CPU by nodegauge


cpu.load_...$CPU Core min/max database utilization onlyDatabase CPU min/maxgauge
clustrix.cpu.load.node.nn.cpu.0load.node.*.cpu.0CPU Core-0 by node database utilization onlyDatabase CPU Core only core-0 by nodegauge
clustrix.cpu.load.node.nn.cpu.nnload.node.*.cpu.[^0]\d*CPU Cores excluding core-0 by node database utilization onlyDatabase CPU core excluding core-0 by nodegauge
clustrix.lockman.wait_ms.(min|max|avg)lockman.wait_ms....$Total time all queries spent waiting on locksTime waiting on query locksgauge
clustrix.qpsqps$Cluster wide Queries per secondCluster QPSgauge
clustrix.response_time.read_usresponse_time.read_usCluster query read latency, number of queries divided by runtimeCluster read latencygauge
clustrix.response_time.write_usresponse_time.write_usCluster query write latency, number of queries divided by runtimeCluster write latencygauge
clustrix.stats.executing_sessionsexecuting_sessionsThe number of sessions currently executingSessions currently executinggauge
clustrix.stats.qps_total.node.nnqps_total.nodeQueries per second by nodeQPS by nodecounter
clustrix.stats.tm_cpu_waittime_ustm_cpu_waittimeThe total amount of time queries spent waiting due to the fair schedulerFair Scheduler wait timecounter
clustrix.stats.tps_total.node.nntps_total.nodeTransactions per second by nodeTPS by nodecounter
clustrix.tpstps$Cluster wide Transactions per secondCluster TPSgauge


Metric NameREGEXPDescriptionMeasuresGraph as
clustrix.stats.Com_(delete|insert|select|update)Com_(delete|insert|select|update)$Number of queries separated by query typeQueries by typecounter
clustrix.stats.Com_(delete|insert|select|update)_usCom_(delete|insert|select|update)_usAmount of run time by query typeTime per query typegauge
clustrix.stats.runtime_read_usruntime_read_usTotal runtime of read queriesRuntime of read queriescounter
clustrix.stats.runtime_write_usruntime_write_usTotal runtime of write queriesRuntime of write queriescounter


Metric NameREGEXPDescriptionMeasuresGraph as
clustrix.replication.slave.relay_log_bytes.(slave)relay_log_bytes[^_]Size in bytes of the relay logs for the replication slaveRelay log bytes for all slavesgauge
clustrix.replication.slave.bytes_read.(slave)slave.bytes_readBytes read by the replication slaveBytes read by slavecounter
clustrix.replication.slave.events_read.(slave)events_readNumber of events read by the replication slaveEvents read by slavecounter
clustrix.replication.slave.seconds_behind_master.(slave)seconds_behind_masterSeconds behind master value for all replication slavesSlave seconds behind mastergauge


Metric NameREGEXPDescriptionMeasuresGraph as


rows_read.node..$Number of rows read by node at the container levelRows read by node, containercounter


rows_written.node..$Number of rows written by node at the container levelRows written by node, containercounter
clustrix.containers.rows_(deleted|inserted|replaced)rows_(deleted|inserted|replaced)$Number of rows deleted, inserted and updated at the container levelRows, deleted, inserted, updatedcounter
clustrix.containers.rows_(read|written)containers.rows_(read|written)$Number of rows read and written at the container levelRows read and written, containercounter


rows_deleted$Number of rows deletes at the container levelRows deleted, container levelcounter


rows_inserted$Number of rows inserted at the container levelRows inserted, container levelcounter


containers.rows_read$Number of rows read at the container levelRows read, container levelcounter


rows_replaced$Number of rows updated at the container levelRows updated, container levelcounter


rows_written$Number of rows written at the container levelRows written, container levelcounter
clustrix.stats.tm_rows_outputtm_rows_outputNumber of rows written by queriesRows written, transaction managercounter
clustrix.stats.tm_rows_readtm_rows_readNumber of rows read by queriesRows read, transaction managercounter


Metric NameREGEXPDescriptionMeasuresGraph as
clustrix.capacity.disks.(avg|max|min)_used_percent[avg|max|min]_used_percentThe min/max/avg disk capacity usedPercentage of disks in usegauge segment)

disk.bytes_read.node.Not applicable for virtual environments (i.e. i3s).Bytes read at the kernel levelcounter segment)disk.bytes_written.node.Not applicable for virtual environments (i.e. i3s).Bytes written at the kernel levelcounter segment)disk.pct_utilization.nodeDisk utilization metric for individual physical disks hosting the vdev file (e.g. through an md RAID device)Disk utilizationgauge segment)read_latency_ms.node.Average read latency from IO time over the same periodStorage read latency, kernel level by nodegauge segment)write_latency_ms.node.Average write latency from IO time over the same periodStorage write latency, kernel level by nodegauge$Device1 bytes read, cluster levelDevice1 bytes read, cluster levelcounter*.vdev.3Device1 bytes read by nodeDevice1 bytes read by nodecounter bytes read per secondDevice1 bytes read per secondgauge$Device1 bytes written, cluster levelDevice1 bytes written, cluster levelcounter*.vdev.3Device1 bytes written by nodeDevice1 bytes written by nodecounter bytes written per secondDevice1 bytes written per secondgauge*.vdev.3Device1 read latency by nodeDevice1 read latency by nodegauge*.vdev.3Device1 sync latency by nodeDevice1 sync latency by nodegauge*.vdev.3Device1 write latency by nodeDevice1 write latency by nodegauge
  • No labels