Service Instance Metrics
This page describes the metrics used to monitor state of an a9s MariaDB Service Instance. For further information on how to monitor an a9s Service Instance, see the Set up Monitoring section of the Application Developer's documentation.
Metric Groups
info
Some metrics may change between the MariaDB versions, therefore they might differ or not be available based on the MariaDB version that is currently running on the Service Instance. In order to be sure, consult the documentation link provided in the metric item description and confirm the information based on both the MariaDB metric and version.
General Status
General MariaDB status.
| Metric Id | Type | Description |
|---|---|---|
aborted_connects | Integer | Number of aborted client connections. This can be due to the client not calling mysql_close() before exiting, the client sleeping without issuing a request to the server for more seconds than specified by wait_timeout or interactive_timeout, or by the client program ending in the midst of transferring data. documentation |
com_delete | Integer | Number of DELETE commands executed. Differs from Handler_delete, which counts the number of times rows have been deleted from tables. documentation |
com_insert | Integer | Number of INSERT commands executed. documentation |
com_select | Integer | Number of SELECT commands executed. Also includes queries that make use of the query cache. documentation |
com_update | Integer | Number of UPDATE commands executed. documentation |
connection_errors_internal | Integer | Number of refused connections due to internal server errors, for example out of memory errors, or failed thread starts. The global value can be flushed by FLUSH STATUS. documentation |
connection_errors_max_connections | Integer | Number of refused connections due to the max_connections limit being reached. The global value can be flushed by FLUSH STATUS. documentation |
innodb_buffer_pool_pages_total | Integer | Total number of InnoDB buffer pool pages. documentation |
Innodb_buffer_pool_pages_free | Integer | Number of free InnoDB buffer pool pages. documentation |
innodb_buffer_pool_read_requests | Integer | Number of requests to read from the InnoDB buffer pool. documentation |
innodb_buffer_pool_reads | Integer | Number of reads that could not be satisfied by the InnoDB buffer pool and had to be read from disk. documentation |
innodb_data_fsyncs | Integer | Number of InnoDB fsync (sync-to-disk) calls. fsync call frequency can be influenced by the innodb_flush_method configuration option.documentation |
innodb_page_size | Integer | Page size used by InnoDB. Defaults to 16KB, can be compiled with a different value. documentation |
innodb_row_lock_waits | Integer | The number of times operations on InnoDB tables had to wait for a row lock. documentation |
max_used_connections | Integer | Max number of connections ever open at the same time. The global value can be flushed by FLUSH STATUS. documentation |
qcache_hits | Integer | Number of requests served by the query cache. The global value can be flushed by FLUSH STATUS. documentation |
questions | Integer | Number of statements executed by the server, excluding COM_PING, COM_STATISTICS, COM_STMT_PREPARE, COM_STMT_CLOSE, and COM_STMT_RESET statements. Differs from Queries in that it doesn't count statements executed within stored programs. documentation |
slow_queries | Integer | Number of queries which took longer than long_query_time to run. The slow query log does not need to be active for this to be recorded. documentation |
threads_connected | Integer | Number of clients connected to the server. See Handling Too Many Connections. The Threads_connected name is inaccurate when the thread pool is in use, since each client connection does not correspond to a dedicated thread in that case. documentation |
threads_running | Integer | Number of client connections that are actively running a command, and not just sleeping while waiting to receive the next command to execute. Some internal system threads also count towards this status variable if they would show up in the output of the SHOW PROCESSLIST statement.documentation |
wsrep_cluster_size | Integer | Number of nodes currently in the cluster. documentation |
wsrep_cluster_status | Integer | Cluster component status. Possible values are PRIMARY (primary group configuration, quorum present), NON_PRIMARY (non-primary group configuration, quorum lost) or DISCONNECTED (not connected to group, retrying). documentation |
reads | Integer | Sum total of com_select and qcache_hits. |
writes | Integer | Sum total of com_insert, com_update, and com_delete. |
innodb_buffer_pool_pages_utilization | It is the calculation by the formula (innodb_buffer_pool_pages_total - innodb_buffer_pool_pages_free) / innodb_buffer_pool_pages_total. | |
wsrep_cluster_state_uuid | UUID | State of the cluster. If it matches the value in local_state_uuid, the local and cluster nodes are in sync. documentation |
wsrep_gcomm_uuid | UUID | The UUID assigned to the node. documentation |
wsrep_last_committed | Integer | Sequence number of the most recently committed transaction. documentation |
wsrep_replicated | Integer | Total number of writesets replicated to other nodes. documentation |
wsrep_received | Integer | Total number of writesets received from other nodes. documentation |
wsrep_ready | String | Whether or not the Galera wsrep provider is ready. Possible values are ON or OFF. documentation |
wsrep_open_transactions | Integer | The number of locally running transactions which have been registered inside the wsrep provider. documentation |
wsrep_open_connections | Integer | The number of open connection objects inside the wsrep provider. documentation |
wsrep_thread_count | Integer | Total number of wsrep (applier/rollbacker) threads. documentation |
wsrep_local_state_uuid | UUID | The node's UUID state. If it matches the value in cluster_state_uuid, the local and cluster nodes are in sync. documentation |
wsrep_local_state_comment | String | Human-readable explanation of the state. documentation |
wsrep_local_state | Integer | Internal Galera Cluster FSM state number. documentation |
wsrep_commit_window | Float | Average distance between the highest and lowest concurrently committed seqno. documentation |
Metrics
*.mariadb.*.*.*.*.status.aborted_connects
*.mariadb.*.*.*.*.status.com_delete
*.mariadb.*.*.*.*.status.com_insert
*.mariadb.*.*.*.*.status.com_select
*.mariadb.*.*.*.*.status.com_update
*.mariadb.*.*.*.*.status.connection_errors_internal
*.mariadb.*.*.*.*.status.connection_errors_max_connections
*.mariadb.*.*.*.*.status.innodb_buffer_pool_pages_total
*.mariadb.*.*.*.*.status.innodb_buffer_pool_read_requests
*.mariadb.*.*.*.*.status.innodb_buffer_pool_reads
*.mariadb.*.*.*.*.status.innodb_data_fsyncs
*.mariadb.*.*.*.*.status.innodb_page_size
*.mariadb.*.*.*.*.status.innodb_row_lock_waits
*.mariadb.*.*.*.*.status.max_used_connections
*.mariadb.*.*.*.*.status.qcache_hits
*.mariadb.*.*.*.*.status.questions
*.mariadb.*.*.*.*.status.slow_queries
*.mariadb.*.*.*.*.status.threads_connected
*.mariadb.*.*.*.*.status.threads_running
*.mariadb.*.*.*.*.status.wsrep_cluster_size
*.mariadb.*.*.*.*.status.wsrep_cluster_status
*.mariadb.*.*.*.*.status.reads
*.mariadb.*.*.*.*.status.writes
*.mariadb.*.*.*.*.status.innodb_buffer_pool_pages_utilization
*.mariadb.*.*.*.*.status.wsrep_cluster_state_uuid
*.mariadb.*.*.*.*.status.wsrep_gcomm_uuid
*.mariadb.*.*.*.*.status.wsrep_last_committed
*.mariadb.*.*.*.*.status.wsrep_replicated
*.mariadb.*.*.*.*.status.wsrep_received
*.mariadb.*.*.*.*.status.wsrep_ready
*.mariadb.*.*.*.*.status.wsrep_open_transactions
*.mariadb.*.*.*.*.status.wsrep_open_connections
*.mariadb.*.*.*.*.status.wsrep_thread_count
*.mariadb.*.*.*.*.status.wsrep_local_state_uuid
*.mariadb.*.*.*.*.status.wsrep_local_state_comment
*.mariadb.*.*.*.*.status.wsrep_local_state
*.mariadb.*.*.*.*.status.wsrep_commit_window
caution
For a9s MariaDB < 10.6 versions the metric names use mysql instead of the mariadb.For example:
*.mysql.*.*.*.*.status.reads
Sys Schema
The sys schema objects can be used for typical tuning and diagnosis use cases.
| Metric Id | Type | Description |
|---|---|---|
sessions.count | Integer | Amount of sessions based on the query SELECT COUNT(*) AS sessions FROM x$session |
io/file.event_class | Integer | The event class. documentation |
io/file.total | Integer | The total number of occurrences of events in the class. documentation |
io/file.total_latency | Integer | The total wait time of timed occurrences of events in the class. documentation |
io/file.min_latency | Integer | The minimum single wait time of timed occurrences of events in the class. documentation |
io/file.avg_latency | Integer | The average wait time per timed occurrence of events in the class. documentation |
io/file.max_latency | Integer | The maximum single wait time of timed occurrences of events in the class. documentation |
io/table.event_class | Integer | The event class. documentation |
io/table.total | Integer | The event class. documentation |
io/table.total_latency | Integer | The total wait time of timed occurrences of events in the class. documentation |
io/table.min_latency | Integer | The minimum single wait time of timed occurrences of events in the class. documentation |
io/table.avg_latency | Integer | The average wait time per timed occurrence of events in the class. documentation |
io/table.max_latency | Integer | The maximum single wait time of timed occurrences of events in the class. documentation |
lock/table.event_class | Integer | The event class. documentation |
lock/table.total | Integer | The total number of occurrences of events in the class. documentation |
lock/table.total_latency | Integer | The total wait time of timed occurrences of events in the class. documentation |
lock/table.min_latency | Integer | The minimum single wait time of timed occurrences of events in the class. documentation |
lock/table.avg_latency | Integer | The average wait time per timed occurrence of events in the class. documentation |
lock/table.max_latency | Integer | The maximum single wait time of timed occurrences of events in the class. documentation |
Metrics
*.mariadb.*.*.*.*.sys.sessions.count
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/file.event_class
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/file.total
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/file.total_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/file.min_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/file.avg_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/file.max_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/table.event_class
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/table.total
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/table.total_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/table.min_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/table.avg_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/io/table.max_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/lock/table.event_class
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/lock/table.total
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/lock/table.total_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/lock/table.min_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/lock/table.avg_latency
*.mariadb.*.*.*.*.sys.wait_classes_global_by_latency.event_class.wait/lock/table.max_latency
caution
For a9s MariaDB < 10.6 versions the metric names use mysql instead of the mariadb. For example:
*.mysql.*.*.*.*.sys.sessions.count