Skip to main content

Cluster-agent

This page describes metrics gathered by coroot-cluster-agent.

Coroot-cluster-agent is a dedicated tool for collecting cluster-wide telemetry data:

  • It gathers database metrics by discovering databases through Coroot's Service Map and Kubernetes control-plane. Using the credentials provided by Coroot or via Kubernetes annotations, the agent connects to the identified databases such as Postgres, MySQL, Redis, Memcached, and MongoDB, collects database-specific metrics, and sends them to Coroot using the Prometheus Remote Write protocol.
  • When --track-database-changes is enabled, the agent tracks schema and configuration changes in databases. Change events are sent to Coroot as OpenTelemetry log records under the DatabaseChanges service name.
  • The agent can be integrated with AWS to discover RDS and ElastiCache clusters and collect their telemetry data.
  • The agent discovers and scrapes custom metrics from annotated pods.
  • The agent monitors GitOps tooling by reading FluxCD and ArgoCD custom resources through its embedded kube-state-metrics and exposing their state as metrics.
  • The agent monitors database backups of clusters managed by Kubernetes operators, reading their custom resources through the embedded kube-state-metrics: Postgres (CloudNativePG, Percona Operator for PostgreSQL), MySQL (Percona Operator for MySQL based on Percona XtraDB Cluster), and MongoDB (Percona Operator for MongoDB).

Postgres​

pg_up​

  • Description: Whether the Postgres server is reachable or not
  • Type: Gauge
  • Source: The agent checks that a connection to the server is still alive on each scrape

pg_probe_seconds​

  • Description: How long it took to execute an empty SQL query (;) on the server. This metric shows the round-trip time between the agent and the server
  • Type: Gauge
  • Source: The time spent executing db.Ping()

pg_scrape_error​

  • Description: Whether a scrape error occurred
  • Type: Gauge
  • Labels: error, warning

pg_info​

pg_setting​

  • Description: Value of the pg_setting variable
  • Type: Gauge
  • Source: pg_settings. The agent only collects variables of the following types: integer, real and bool
  • Labels: name, unit

pg_connections​

  • Description: The number of the database connections
  • Type: Gauge
  • Source: pg_stat_activity
  • Labels:
    • db
    • user
    • state: current state of the connection, < active | idle | idle in transaction >
    • wait_event_type: type of event that the connection is waiting for.
    • query - If the state of a connection is active, this is the currently executing query. For idle in transaction connections, this is the last executed query. This label holds a normalized and obfuscated query.

pg_autovacuum_workers​

  • Description: Number of running autovacuum worker processes
  • Type: Gauge
  • Source: pg_stat_activity (backend_type = 'autovacuum worker')

pg_latency_seconds​

pg_db_queries_per_second​

  • Description: Number of queries executed in the database
  • Type: Gauge
  • Source: Aggregation of pg_stat_activity.state = 'Active' and pg_stat_statements.calls
  • Labels: db

pg_lock_awaiting_queries​

  • Description: Number of queries awaiting a lock
  • Type: Gauge
  • Source: Number of connections with pg_stat_activity.wait_event_type = 'Lock'. The blocking_query label is calculated using the pg_blocking_pids function
  • Labels: db, user, blocking_query (the query holding the lock)

Query Metrics​

The pg_stat_statements view shows statistics only for queries that have been completed. So, to provide comprehensive statistics, the agent extends this with data about the currently active queries from the pg_stat_activity view.

Collecting stats about each query would produce metrics with very high cardinality. However, the primary purpose of such metrics is to show the most resource-consuming queries. So, the agent collects these metrics only for TOP-20 queries by total execution time.

Each metric described below has query, db and user labels. Query is a normalized and obfuscated query from pg_stat_statements.query, and pg_stat_activity.query.

For example, the following queries:

SELECT * FROM tbl WHERE id='1';
SELECT * FROM tbl WHERE id='2';

will be grouped to

SELECT * FROM tbl WHERE id=?;

pg_top_query_calls_per_second​

  • Description: Number of times the query has been executed
  • Type: Gauge
  • Source: pg_stat_statements.calls and pg_stat_activity.state = 'Active'
  • Labels: db, user, query

pg_top_query_time_per_second​

  • Description: Time spent executing the query
  • Type: Gauge
  • Source: clock_timestamp()-pg_stat_activity.query_start and pg_stat_statements.total_time
  • Labels: db, user, query

pg_top_query_io_time_per_second​

  • Description: Time the query spent awaiting I/O
  • Type: Gauge
  • Source: pg_stat_activity.wait_event_type = 'IO', pg_stat_statements.blk_read_time and pg_stat_statements.blk_write_time
  • Labels: db, user, query

Replication metrics​

pg_wal_receiver_status​

  • Description: WAL receiver status: 1 if the receiver is connected, otherwise 0
  • Type: Gauge
  • Source: pg_stat_wal_receiver and pg_settings[primary_conninfo]
  • Labels: sender_host, sender_port

pg_wal_replay_paused​

  • Description: Whether WAL replay paused or not
  • Type: Gauge
  • Source: pg_is_wal_replay_paused() or pg_is_xlog_replay_paused()

pg_wal_current_lsn​

  • Description: Current WAL sequence number
  • Type: Counter
  • Source: pg_current_wal_lsn() or pg_current_xlog_location()

pg_wal_receive_lsn​

  • Description: WAL sequence number that has been received and synced to disk by streaming replication.
  • Type: Counter
  • Source: pg_last_wal_receive_lsn() or pg_last_xlog_receive_location()

pg_wal_reply_lsn​

  • Description: WAL sequence number that has been replayed during recovery
  • Type: Counter
  • Source: pg_last_wal_replay_lsn() or pg_last_xlog_replay_location()

WAL size and archiving​

pg_wal_size_bytes​

  • Description: Size of the WAL directory
  • Type: Gauge
  • Source: pg_ls_waldir()

pg_replication_slot_retained_wal_bytes​

  • Description: Amount of WAL retained for the replication slot
  • Type: Gauge
  • Source: pg_replication_slots (restart_lsn); wal_status is reported on Postgres >= 13
  • Labels: slot, active, wal_status

pg_wal_archived_segments_total​

  • Description: Number of WAL files successfully archived
  • Type: Counter
  • Source: pg_stat_archiver

pg_wal_archive_failures_total​

  • Description: Number of failed attempts to archive WAL files
  • Type: Counter
  • Source: pg_stat_archiver

pg_wal_archiving_status​

  • Description: 1 if the last WAL archive attempt succeeded, 0 if it failed
  • Type: Gauge
  • Source: pg_stat_archiver

Checkpoint metrics​

The agent tracks checkpoint activity from pg_stat_checkpointer (Postgres >= 17) or pg_stat_bgwriter (older versions). The counters are rebased and accumulated over the agent's lifetime.

pg_checkpoints_scheduled_total​

  • Description: Number of scheduled checkpoints, including skipped ones
  • Type: Counter
  • Source: pg_stat_checkpointer (Postgres >= 17) or pg_stat_bgwriter
  • Labels: type (timed, requested)

pg_checkpoints_total​

  • Description: Number of checkpoints that have been completed
  • Type: Counter
  • Source: pg_stat_checkpointer (Postgres >= 17) or pg_stat_bgwriter

pg_restartpoints_total​

  • Description: Number of restartpoints that have been completed on a standby
  • Type: Counter
  • Source: pg_stat_checkpointer.restartpoints_done (Postgres >= 17)

pg_buffers_written_total​

  • Description: Total number of dirty buffers flushed to disk
  • Type: Counter
  • Source: pg_stat_checkpointer (Postgres >= 17) or pg_stat_bgwriter
  • Labels: source (checkpointer)

pg_time_since_last_checkpoint_seconds​

  • Description: Seconds since the last checkpoint observed by the agent
  • Type: Gauge
  • Source: Measured by the agent from the checkpoint completion counter

pg_wal_since_last_checkpoint_bytes​

  • Description: Amount of WAL written since the last completed checkpoint (to be replayed in the case of a crash)
  • Type: Gauge
  • Source: pg_control_checkpoint() (redo_lsn)

Transaction ID age​

pg_xid_age​

  • Description: Transactions since the oldest unfrozen transaction ID (age of datfrozenxid)
  • Type: Gauge
  • Source: pg_database
  • Labels: db

pg_multixact_age​

  • Description: Multixacts since the oldest unfrozen multixact ID (age of datminmxid)
  • Type: Gauge
  • Source: pg_database
  • Labels: db

pg_oldest_xmin_age​

  • Description: Age, in transactions, of the oldest transaction ID held back from freezing, by holder
  • Type: Gauge
  • Source: pg_stat_activity, pg_replication_slots, pg_prepared_xacts (Postgres >= 10)
  • Labels: holder (running_transaction, standby_feedback, replication_slot, prepared_transaction)

pg_transaction_seconds​

  • Description: Age of the longest-running transaction, by query

  • Type: Gauge

  • Source: pg_stat_activity (now() - xact_start)

  • Labels:

    • db
    • user
    • query - the statement the transaction is running. For an idle in transaction session this is the last statement it executed, which is the code path that opened the transaction. This label holds a normalized and obfuscated query, and is ~empty when obfuscation leaves no text (for example a statement that is only a comment).

    Where pg_oldest_xmin_age says that a running transaction is holding the vacuum horizon back, this says which one, so a report can name the statement instead of leaving you to find it in pg_stat_activity by hand. That matters because the database named in a wraparound or bloat report is whichever one holds the oldest datfrozenxid, which is often not the database the transaction is running in.

    Only transactions open for at least 10 seconds are reported, and at most the 20 longest query shapes.

Change tracking​

When --track-database-changes is enabled, the agent detects and emits change events for:

  • Schema changes — The agent periodically snapshots the DDL of every table (columns, constraints, indexes) across all databases. When a table's DDL changes between consecutive snapshots, a change event is emitted with a unified diff. Only schema modifications are tracked (e.g., ALTER TABLE, index creation/removal). The snapshot is collected by connecting to each database and querying pg_catalog and information_schema.
  • Settings changes — The agent snapshots all pg_settings values each cycle. When a setting changes (e.g., after a configuration reload or restart), a change event is emitted with the diff. Session-level and client-level overrides are excluded.

Each change event includes db.system, db.target, db.name, db_change.object, and db_change.type attributes.

Size and bloat metrics​

The agent collects database and table size metrics. For table sizes, only the top 20 largest tables across all databases are reported.

pg_database_size_bytes​

  • Description: Total size of the database in bytes
  • Type: Gauge
  • Source: pg_database_size()
  • Labels: db

pg_table_size_bytes​

  • Description: Total size of the table in bytes including indexes and TOAST
  • Type: Gauge
  • Source: pg_total_relation_size()
  • Labels: db, schema, table

pg_table_size_growth_bytes_per_second​

  • Description: Table size growth rate in bytes per second. Only the top 20 fastest growing tables across all databases are reported. Requires at least two collection cycles to compute.
  • Type: Gauge
  • Source: Computed from consecutive pg_total_relation_size() measurements
  • Labels: db, schema, table

When --track-database-bloat is enabled, the agent also estimates wasted space (bloat) for tables and indexes from planner statistics (pg_class.reltuples/relpages and pg_stats column widths), without scanning table data. TOAST relations are excluded, and only the top tables and indexes by estimated bloat are reported per database. Estimates are approximate and depend on up-to-date ANALYZE/autovacuum statistics.

pg_db_table_bloat_bytes​

  • Description: Estimated wasted space across all tables of the database
  • Type: Gauge
  • Source: Estimated from pg_class and pg_stats
  • Labels: db

pg_db_index_bloat_bytes​

  • Description: Estimated wasted space across all indexes of the database
  • Type: Gauge
  • Source: Estimated from pg_class and pg_stats
  • Labels: db

pg_table_bloat_bytes​

  • Description: Estimated wasted space in the table heap
  • Type: Gauge
  • Source: Estimated from pg_class and pg_stats
  • Labels: db, schema, table

pg_index_bloat_bytes​

  • Description: Estimated wasted space in the index
  • Type: Gauge
  • Source: Estimated from pg_class and pg_stats
  • Labels: db, schema, table, index

When --track-database-sizes is enabled, the agent also reports dead-row statistics — a leading indicator of autovacuum falling behind, distinct from bloat. The dead/live counts let Coroot compute autovacuum pressure (how many times past its own autovacuum trigger a table sits), and dead bytes provides materiality. All three are emitted from one query with one top-N ranking (by dead bytes), so every reported table carries the complete set.

pg_table_dead_tuple_bytes​

  • Description: Estimated size of dead tuples not yet reclaimed by vacuum (heap size × dead fraction)
  • Type: Gauge
  • Source: pg_stat_user_tables (n_dead_tup, n_live_tup) and pg_relation_size()
  • Labels: db, schema, table

pg_table_dead_tuples​

  • Description: Number of dead tuples not yet reclaimed by vacuum
  • Type: Gauge
  • Source: pg_stat_user_tables (n_dead_tup)
  • Labels: db, schema, table

pg_table_live_tuples​

  • Description: Estimated number of live tuples
  • Type: Gauge
  • Source: pg_stat_user_tables (n_live_tup)
  • Labels: db, schema, table

pg_table_seconds_since_last_autovacuum​

  • Description: Seconds since the last autovacuum of the table. Not reported for tables that have never been autovacuumed.
  • Type: Gauge
  • Source: pg_stat_user_tables (now() - last_autovacuum)
  • Labels: db, schema, table

pg_table_setting​

  • Description: Info metric (value always 1) carrying a table's per-table autovacuum and autoanalyze reloption overrides as labels. Reported only for tables that override at least one setting; an unset override is an empty label. Coroot uses the trigger overrides to compute pressure against the table's own vacuum and analyze triggers, and the cost overrides to pinpoint per-table throttling.
  • Type: Gauge
  • Source: pg_class.reloptions
  • Labels: db, schema, table, and the override values: autovacuum_disabled (1 when autovacuum_enabled=false, which disables autoanalyze too), autovacuum_vacuum_scale_factor, autovacuum_vacuum_threshold, autovacuum_vacuum_cost_delay, autovacuum_vacuum_cost_limit, autovacuum_analyze_scale_factor, autovacuum_analyze_threshold

pg_table_vacuum_in_progress​

  • Description: 1 if a vacuum is currently running on the table. Reported only while a vacuum is in progress. Lets Coroot tell a table that has a worker on it (possibly crawling) from one waiting for a free worker.
  • Type: Gauge
  • Source: pg_stat_progress_vacuum (Postgres >= 9.6)
  • Labels: db, schema, table

pg_table_vacuum_throttled​

  • Description: 1 if the running vacuum is sleeping on the cost-based delay (VacuumDelay wait event) at scrape time, 0 otherwise. Reported only while a vacuum is in progress; a high average means the vacuum is being throttled by autovacuum_vacuum_cost_delay/autovacuum_vacuum_cost_limit.
  • Type: Gauge
  • Source: pg_stat_progress_vacuum joined with pg_stat_activity (wait_event = 'VacuumDelay')
  • Labels: db, schema, table

pg_table_mods_since_analyze​

  • Description: Rows modified (inserted, updated, or deleted) since the table's planner statistics were last analyzed. Unlike dead tuples this includes inserts, so it is collected as its own top-N (ranked by analyze pressure) rather than reusing the dead-tuple set. Coroot divides it by the autoanalyze trigger to compute how stale the statistics are.
  • Type: Gauge
  • Source: pg_stat_user_tables (n_mod_since_analyze)
  • Labels: db, schema, table

pg_table_reltuples​

  • Description: The planner's estimate of the number of live rows in the table. Used as the denominator of analyze/vacuum pressure. Unlike n_live_tup it is persisted in pg_class, so it survives a replica promotion (which resets the cumulative pg_stat_* counters) and keeps pressure from spiking to a false value on a freshly promoted primary.
  • Type: Gauge
  • Source: pg_class.reltuples
  • Labels: db, schema, table

pg_table_seconds_since_last_analyze​

  • Description: Seconds since the table's planner statistics were last refreshed by ANALYZE or autoanalyze. Not reported for tables that have never been analyzed.
  • Type: Gauge
  • Source: pg_stat_user_tables (now() - greatest(last_analyze, last_autoanalyze))
  • Labels: db, schema, table

Postgres backups​

Backup state is collected through the agent's embedded kube-state-metrics from the custom resources of CloudNativePG (cnpg) and the Percona Operator for PostgreSQL (pgBackRest, percona). Every metric carries an operator label identifying the source and, via the common namespace/name labels, correlates to the corresponding DatabaseCluster application in Coroot, so a single operator-agnostic set of metrics powers the backup inspection.

pg_backup_target_info​

  • Description: A configured backup destination. cnpg reports a single object-storage target, and pgBackRest reports one series per repository. Coroot assembles the destination from the explicit path or the object-storage sub-fields.
  • Type: Info
  • Source: Cluster.spec.backup (cnpg), PerconaPGCluster.spec.backups.pgbackrest.repos (Percona)
  • Labels: operator, method, path, endpoint, s3_bucket, s3_endpoint, gcs_bucket, azure_container, schedule, retention_policy

pg_cluster_status​

  • Description: A cluster status condition (e.g. ReadyForBackup, LastBackupSucceeded, ContinuousArchiving, PGBackRestRepoHostReady). Value is 1 for the currently-active series. The reason feeds Coroot's "why backups are failing" hint.
  • Type: Info
  • Source: .status.conditions
  • Labels: operator, type, status, reason

pg_backup_last_successful_timestamp_seconds​

  • Description: Time of the last successful backup, per method. Reported by cnpg. For Percona it is derived from the individual backup runs.
  • Type: Gauge
  • Source: Cluster.status.lastSuccessfulBackupByMethod
  • Labels: operator, method

pg_backup_first_recoverability_point_timestamp_seconds​

  • Description: The oldest point in time the cluster can be restored to (start of the recovery window), per method.
  • Type: Gauge
  • Source: Cluster.status.firstRecoverabilityPointByMethod
  • Labels: operator, method

pg_backup_last_failed_timestamp_seconds​

  • Description: Time of the last failed backup.
  • Type: Gauge
  • Source: Cluster.status.lastFailedBackup
  • Labels: operator

pg_backup_schedule_info​

  • Description: The backup schedule (cron).
  • Type: Info
  • Source: ScheduledBackup.spec.schedule (cnpg)
  • Labels: operator, cluster, schedule

pg_backup_next_scheduled_timestamp_seconds​

  • Description: When the next scheduled backup is due. Coroot also derives an expected next run from the schedule and the last backup, so an overdue schedule is detected even when the operator stops advancing this value.
  • Type: Gauge
  • Source: ScheduledBackup.status.nextScheduleTime (cnpg)
  • Labels: operator, cluster

pg_backup_info​

  • Description: An individual backup run (one series per backup object), used to list recent backups. Carries immutable identity only. The run's phase is in pg_backup_status.
  • Type: Info
  • Source: Backup (cnpg), PerconaPGBackup (Percona)
  • Labels: operator, cluster, method, kind, path

pg_backup_status​

  • Description: The current phase of a backup run (e.g. Running, Succeeded, Failed, completed). Value is 1 for the currently-active series (a run's phase changes over its lifecycle, so only the live series is used).
  • Type: Info
  • Source: Backup.status.phase (cnpg), PerconaPGBackup.status.state (Percona)
  • Labels: operator, status

pg_backup_completed_timestamp_seconds​

  • Description: When a backup run completed.
  • Type: Gauge
  • Source: Backup.status.stoppedAt (cnpg), PerconaPGBackup.status.completed (Percona)
  • Labels: operator, cluster

MySQL​

mysql_up​

  • Description: Whether the MySQL server is reachable or not
  • Type: Gauge

mysql_scrape_error​

  • Description: Whether a scrape error occurred
  • Type: Gauge
  • Labels: error, warning

mysql_info​

  • Description: The server info
  • Type: Gauge
  • Labels: server_version, server_id, server_uuid

mysql_top_query_calls_per_second​

  • Description: Number of times the query has been executed
  • Type: Gauge
  • Source: performance_schema.events_statements_summary_by_digest
  • Labels: schema, query

mysql_top_query_time_per_second​

  • Description: Time spent executing the query
  • Type: Gauge
  • Source: performance_schema.events_statements_summary_by_digest
  • Labels: schema, query

mysql_top_query_lock_time_per_second​

  • Description: Time the query spent waiting for locks
  • Type: Gauge
  • Source: performance_schema.events_statements_summary_by_digest
  • Labels: schema, query

mysql_top_query_rows_examined_per_second​

  • Description: Rows examined per second by the query (vs rows returned: a high ratio means a missing index)
  • Type: Gauge
  • Source: performance_schema.events_statements_summary_by_digest
  • Labels: schema, query

mysql_top_query_rows_sent_per_second​

  • Description: Rows returned per second by the query
  • Type: Gauge
  • Source: performance_schema.events_statements_summary_by_digest
  • Labels: schema, query

Lock waits​

The agent samples performance_schema.data_lock_waits on each scrape and reports the current number of queries blocked on row locks. Unlike mysql_top_query_lock_time_per_second (which is derived from cumulative counters and only updates once a blocked query completes), these gauges reflect the live contention at scrape time, which makes them suitable for correlation with client latency. Waiters are deduplicated by requesting transaction id, so a query blocked by several holders is counted once. Not collected on MariaDB.

mysql_locked_queries​

  • Description: Number of queries currently waiting for a lock
  • Type: Gauge
  • Source: performance_schema.data_lock_waits joined to performance_schema.threads and performance_schema.events_statements_current on the requesting thread. The query label is the obfuscated statement digest of the waiting (victim) query.
  • Labels: schema, query (the query waiting for the lock)

mysql_lock_awaiting_queries​

  • Description: Number of queries currently awaiting a lock, grouped by the query holding it
  • Type: Gauge
  • Source: performance_schema.data_lock_waits joined to performance_schema.threads and performance_schema.events_statements_current on the blocking thread. The blocking_query label is the obfuscated statement digest of the query holding the lock.
  • Labels: schema, blocking_query (the query holding the lock)

Replication metrics​

mysql_replication_io_status​

  • Description: Whether the replication IO thread is running
  • Type: Gauge
  • Labels: source_server_id, source_server_uuid, state, last_error

mysql_replication_sql_status​

  • Description: Whether the replication SQL thread is running
  • Type: Gauge
  • Labels: source_server_id, source_server_uuid, state, last_error

mysql_replication_lag_seconds​

  • Description: Seconds behind master
  • Type: Gauge
  • Labels: source_server_id, source_server_uuid

Connection metrics​

mysql_connections_max​

  • Description: Maximum number of allowed connections
  • Type: Gauge
  • Source: SHOW GLOBAL VARIABLES (max_connections)

mysql_connections_current​

  • Description: Current number of connected threads
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS (Threads_connected)

mysql_connections_total​

  • Description: Total number of connections since server start
  • Type: Counter
  • Source: SHOW GLOBAL STATUS (Connections)

mysql_connections_aborted_total​

  • Description: Total number of aborted connection attempts
  • Type: Counter
  • Source: SHOW GLOBAL STATUS (Aborted_connects)

mysql_connection_errors_max_connections_total​

  • Description: Number of connections refused because max_connections was reached
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

Traffic metrics​

mysql_traffic_received_bytes_total​

  • Description: Total bytes received by the server
  • Type: Counter
  • Source: SHOW GLOBAL STATUS (Bytes_received)

mysql_traffic_sent_bytes_total​

  • Description: Total bytes sent by the server
  • Type: Counter
  • Source: SHOW GLOBAL STATUS (Bytes_sent)

mysql_queries_total​

  • Description: Total number of queries executed
  • Type: Counter
  • Source: SHOW GLOBAL STATUS (Questions)

mysql_slow_queries_total​

  • Description: Total number of slow queries
  • Type: Counter
  • Source: SHOW GLOBAL STATUS (Slow_queries)

mysql_top_table_io_wait_time_per_second​

  • Description: Time spent on table I/O operations
  • Type: Gauge
  • Source: performance_schema.table_io_waits_summary_by_table
  • Labels: schema, table, operation

Change tracking​

When --track-database-changes is enabled, the agent detects and emits change events for:

  • Schema changes — The agent periodically snapshots the DDL of every table (columns, indexes, foreign keys) across all databases by querying information_schema. When a table's DDL changes between consecutive snapshots, a change event is emitted with a unified diff.
  • Settings changes — The agent snapshots all SHOW GLOBAL VARIABLES values each cycle. When a variable changes (e.g., after a SET GLOBAL or restart), a change event is emitted with the diff.

Each change event includes db.system, db.target, db.name, db_change.object, and db_change.type attributes.

Size metrics​

The agent collects database and table size metrics from information_schema.tables. For table sizes, only the top 20 largest tables across all databases are reported.

mysql_database_size_bytes​

  • Description: Total size of the database in bytes, computed as the sum of data_length + index_length for all tables
  • Type: Gauge
  • Source: information_schema.tables
  • Labels: db

mysql_table_size_bytes​

  • Description: Total size of the table in bytes (data_length + index_length)
  • Type: Gauge
  • Source: information_schema.tables
  • Labels: db, table

mysql_table_size_growth_bytes_per_second​

  • Description: Table size growth rate in bytes per second. Only the top 20 fastest growing tables across all databases are reported. Requires at least two collection cycles to compute.
  • Type: Gauge
  • Source: Computed from consecutive information_schema.tables measurements
  • Labels: db, table

Server load metrics​

mysql_threads_running​

  • Description: Number of threads that are not sleeping
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_created_tmp_disk_tables_total​

  • Description: Number of internal on-disk temporary tables created (sorts/joins spilling to disk)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

InnoDB buffer pool​

The buffer pool is InnoDB's page cache. Page counts are reported as-is; multiply by mysql_innodb_page_size_bytes to get bytes. The hit rate is 1 - mysql_innodb_buffer_pool_reads_total / mysql_innodb_buffer_pool_read_requests_total: a falling hit rate means the working set no longer fits and reads are going to disk.

mysql_innodb_buffer_pool_read_requests_total​

  • Description: Logical read requests to the InnoDB buffer pool
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_reads_total​

  • Description: Read requests that could not be satisfied from the buffer pool and went to disk
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_write_requests_total​

  • Description: Writes done to the InnoDB buffer pool
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_pages_total​

  • Description: Total pages in the InnoDB buffer pool
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_pages_free​

  • Description: Free pages in the InnoDB buffer pool
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_pages_dirty​

  • Description: Dirty (modified, not yet flushed) pages in the buffer pool
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_pages_data​

  • Description: Pages holding data (clean + dirty) in the buffer pool
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_wait_free_total​

  • Description: Times a write had to wait for a free buffer-pool page (buffer pool pressure)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_buffer_pool_pages_flushed_total​

  • Description: Buffer-pool pages flushed to disk
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_page_size_bytes​

  • Description: InnoDB page size in bytes (converts buffer-pool page counts to bytes)
  • Type: Gauge
  • Source: SHOW GLOBAL VARIABLES

InnoDB row operations and transactions​

mysql_innodb_rows_read_total​

  • Description: Rows read by InnoDB
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_rows_inserted_total​

  • Description: Rows inserted by InnoDB
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_rows_updated_total​

  • Description: Rows updated by InnoDB
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_rows_deleted_total​

  • Description: Rows deleted by InnoDB
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_commit_total​

  • Description: COMMIT statements executed
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_rollback_total​

  • Description: ROLLBACK statements executed
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_sort_merge_passes_total​

  • Description: Merge passes for sorts that spilled to disk (sort_buffer_size too small)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

InnoDB I/O and redo log​

mysql_innodb_data_reads_total​

  • Description: InnoDB data read operations (OS reads)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_data_writes_total​

  • Description: InnoDB data write operations (OS writes)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_data_read_bytes_total​

  • Description: Bytes read by InnoDB
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_data_written_bytes_total​

  • Description: Bytes written by InnoDB
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_data_fsyncs_total​

  • Description: InnoDB fsync() operations
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_log_waits_total​

  • Description: Times a write had to wait for the redo log buffer to be flushed
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_os_log_written_bytes_total​

  • Description: Bytes written to the InnoDB redo log
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

Lock and contention metrics​

Row-lock counters come from SHOW GLOBAL STATUS. Deadlocks, lock-wait timeouts and the undo history list are not exposed there, so they are read from information_schema.INNODB_METRICS (requires the PROCESS privilege). A growing history list means purge is lagging behind a long-running transaction.

mysql_innodb_row_lock_waits_total​

  • Description: Number of times a row lock had to be waited for
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_row_lock_time_seconds_total​

  • Description: Total time spent waiting for InnoDB row locks, seconds
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_row_lock_current_waits​

  • Description: Row locks currently being waited for
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_table_locks_waited_total​

  • Description: Table-lock requests that had to wait (LOCK TABLES, MyISAM, DDL metadata locks)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_table_locks_immediate_total​

  • Description: Table-lock requests granted immediately
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_innodb_deadlocks_total​

  • Description: Transactions rolled back by InnoDB deadlocks
  • Type: Counter
  • Source: information_schema.INNODB_METRICS

mysql_innodb_lock_wait_timeouts_total​

  • Description: Transactions rolled back after waiting innodb_lock_wait_timeout for a row lock
  • Type: Counter
  • Source: information_schema.INNODB_METRICS

mysql_innodb_history_list_length​

  • Description: Undo records not yet purged (history list length); grows while a long transaction holds them
  • Type: Gauge
  • Source: information_schema.INNODB_METRICS

mysql_innodb_transaction_seconds​

  • Description: Age of the longest-running active InnoDB transactions, by query shape

  • Type: Gauge

  • Source: information_schema.innodb_trx

  • Labels: query - a normalized and obfuscated query. InnoDB reports no query for a transaction that is idle between statements, so those are labelled (idle in transaction).

    Only transactions open for at least 10 seconds are reported, and at most the 20 longest query shapes. The Postgres equivalent is pg_transaction_seconds.

Galera metrics​

Collected when the instance is part of a Galera cluster (Percona XtraDB Cluster, MariaDB Galera), detected by the presence of wsrep_cluster_size in SHOW GLOBAL STATUS. Flow control is reported as a counter of seconds, so rate() gives the fraction of time writes were paused.

mysql_wsrep_cluster_size​

  • Description: Number of nodes in the Galera cluster component
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_cluster_status​

  • Description: Status of the cluster component this node is in (Primary means it has quorum)
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS
  • Labels: status

mysql_wsrep_local_state​

  • Description: Current Galera node state (1), with the state name in the label
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS
  • Labels: state

mysql_wsrep_ready​

  • Description: Whether the node can accept queries (1 = ON)
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_connected​

  • Description: Whether the node is connected to the cluster (1 = ON)
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_flow_control_paused_seconds_total​

  • Description: Total time writes were paused by flow control
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_local_recv_queue​

  • Description: Current number of write-sets waiting to be applied on this node
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_local_send_queue​

  • Description: Current number of write-sets waiting to be sent from this node
  • Type: Gauge
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_local_cert_failures_total​

  • Description: Number of transactions that failed certification (write conflicts)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

mysql_wsrep_local_bf_aborts_total​

  • Description: Number of local transactions aborted by replication (brute-force aborts)
  • Type: Counter
  • Source: SHOW GLOBAL STATUS

Group Replication metrics​

Collected when group_replication_group_name is set.

mysql_group_replication_member_state​

  • Description: Group Replication state of this member (1), with the state name in the label
  • Type: Gauge
  • Source: performance_schema.replication_group_members
  • Labels: state

mysql_group_replication_cluster_size​

  • Description: Number of members in the Group Replication group
  • Type: Gauge
  • Source: performance_schema.replication_group_members

mysql_group_replication_members_online​

  • Description: Number of members currently ONLINE in the group
  • Type: Gauge
  • Source: performance_schema.replication_group_members

mysql_group_replication_transactions_in_queue​

  • Description: Transactions waiting in the certification queue on this member
  • Type: Gauge
  • Source: performance_schema.replication_group_member_stats

mysql_group_replication_transactions_remote_in_applier_queue​

  • Description: Remote transactions waiting to be applied on this member
  • Type: Gauge
  • Source: performance_schema.replication_group_member_stats

mysql_group_replication_conflicts_detected_total​

  • Description: Number of transactions that failed certification (conflicts) on this member
  • Type: Counter
  • Source: performance_schema.replication_group_member_stats

Binary log and undo metrics​

Binary logs and undo tablespaces grow independently of table data and are a common cause of a full data volume. If binary logging is disabled the binary log query is skipped.

mysql_binlog_size_bytes​

  • Description: Total size of the binary logs on disk
  • Type: Gauge
  • Source: SHOW BINARY LOGS

mysql_binlog_files​

  • Description: Number of binary log files on disk
  • Type: Gauge
  • Source: SHOW BINARY LOGS

mysql_binlog_expire_seconds​

  • Description: binlog_expire_logs_seconds; 0 means binary logs are never purged automatically
  • Type: Gauge
  • Source: SHOW GLOBAL VARIABLES (binlog_expire_logs_seconds, or expire_logs_days on older MySQL and MariaDB)

mysql_undo_size_bytes​

  • Description: Total size of the InnoDB undo tablespaces on disk
  • Type: Gauge
  • Source: information_schema.INNODB_TABLESPACES (INNODB_SYS_TABLESPACES on MariaDB)

MySQL backups​

When running on Kubernetes, backup state is collected through the agent's embedded kube-state-metrics from the custom resources of the Percona Operator for MySQL based on Percona XtraDB Cluster (PerconaXtraDBCluster, PerconaXtraDBClusterBackup). Every metric carries an operator label (percona) and, via the common namespace/name labels, correlates to the corresponding DatabaseCluster application in Coroot.

mysql_backup_target_info​

  • Description: A configured backup storage, one series per entry of spec.backup.storages (the method label is the storage name). Coroot assembles the destination from the S3 bucket and prefix or the Azure container.
  • Type: Info
  • Source: PerconaXtraDBCluster.spec.backup.storages
  • Labels: operator, method, type, s3_bucket, s3_endpoint, s3_prefix, azure_container

mysql_backup_schedule_info​

  • Description: A scheduled backup (one series per entry of spec.backup.schedule). Coroot uses the schedule to detect overdue backups; method is the storage the schedule writes to.
  • Type: Info
  • Source: PerconaXtraDBCluster.spec.backup.schedule
  • Labels: operator, task, schedule, method

mysql_backup_pitr_info​

  • Description: Whether point-in-time recovery (continuous binlog upload) is enabled for the cluster.
  • Type: Info
  • Source: PerconaXtraDBCluster.spec.backup.pitr
  • Labels: operator, enabled

mysql_cluster_status​

  • Description: The cluster state reported by the operator (e.g. ready, initializing, error). Value is 1 for the currently-active series.
  • Type: Info
  • Source: PerconaXtraDBCluster.status.state
  • Labels: operator, status

mysql_backup_info​

  • Description: An individual backup run (one series per PerconaXtraDBClusterBackup object), used to list recent backups. Carries immutable identity only; the run's phase is in mysql_backup_status. cluster is the name of the PerconaXtraDBCluster the backup belongs to.
  • Type: Info
  • Source: PerconaXtraDBClusterBackup.spec (pxcCluster, storageName), .status (storage_type, destination)
  • Labels: operator, cluster, method, kind, path

mysql_backup_status​

  • Description: The current phase of a backup run (e.g. Starting, Running, Succeeded, Failed). Value is 1 for the currently-active series (a run's phase changes over its lifecycle, so only the live series is used).
  • Type: Info
  • Source: PerconaXtraDBClusterBackup.status.state
  • Labels: operator, status

mysql_backup_completed_timestamp_seconds​

  • Description: When a backup run completed.
  • Type: Gauge
  • Source: PerconaXtraDBClusterBackup.status.completed
  • Labels: operator

MongoDB​

mongo_up​

  • Description: Whether the MongoDB server is reachable or not
  • Type: Gauge

mongo_scrape_error​

  • Description: Whether a scrape error occurred
  • Type: Gauge
  • Labels: error, warning

mongo_info​

  • Description: The server info; flavor is percona for Percona Server for MongoDB, mongodb otherwise
  • Type: Gauge
  • Labels: server_version, flavor

mongo_rs_status​

  • Description: Replica set status: 1 if the member is part of a replica set. The member reports its own role (role label), which Coroot uses to identify the primary and to derive each secondary's replication lag from mongo_rs_last_applied_timestamp_ms.
  • Type: Gauge
  • Source: replSetGetStatus (the self member)
  • Labels: rs, role

mongo_rs_last_applied_timestamp_ms​

  • Description: Timestamp of the member's last applied operation, in milliseconds. Coroot computes replication lag as the primary's value minus each secondary's.
  • Type: Gauge
  • Source: replSetGetStatus (optimes.appliedOpTime)

mongo_rs_member_config_info​

  • Description: Replica set member configuration
  • Type: Gauge
  • Source: replSetGetConfig
  • Labels: rs, member, arbiter, votes

mongo_rs_config_info​

  • Description: Replica set configuration; Coroot warns when write_concern_majority_journal_default is false (acknowledged majority writes may be lost if a majority of members crash simultaneously)
  • Type: Gauge
  • Source: replSetGetConfig
  • Labels: rs, write_concern_majority_journal_default

mongo_profiling_level​

  • Description: Database profiling level (0 - off, 1 - slow operations, 2 - all operations). Coroot suggests enabling profiling when it is off everywhere, since it is the source of per-query statistics.
  • Type: Gauge
  • Source: the profile command (get-only, profile: -1)
  • Labels: db

mongo_oplog_window_seconds​

  • Description: Time span between the oldest and the newest oplog entries
  • Type: Gauge
  • Source: local.oplog.rs

mongo_oplog_size_bytes / mongo_oplog_max_size_bytes​

  • Description: Current size of the oplog data and the configured maximum
  • Type: Gauge
  • Source: collStats on local.oplog.rs

mongo_connections_current / mongo_connections_active / mongo_connections_max​

  • Description: Number of client connections, connections currently executing operations, and the connection limit
  • Type: Gauge
  • Source: serverStatus.connections

mongo_connections_created_total​

  • Description: Total number of connections created
  • Type: Counter

mongo_connections_rejected_total​

  • Description: Total number of rejected connections
  • Type: Counter

mongo_connections_by_app​

  • Description: Number of connections by client application name (top 20)
  • Type: Gauge
  • Source: $currentOp
  • Labels: app

mongo_opcounters_total​

  • Description: Total number of operations by type (insert, query, update, delete, getmore, command)
  • Type: Counter
  • Source: serverStatus.opcounters
  • Labels: op

mongo_documents_returned_total​

  • Description: Total number of documents returned by queries; compared with the scanned counters to detect inefficient queries
  • Type: Counter
  • Source: serverStatus.metrics.document

mongo_op_latency_seconds_total / mongo_op_latency_ops_total​

  • Description: Total operation latency and operation count by type (read, write, command). The average latency is rate(latency) / rate(ops).
  • Type: Counter
  • Source: serverStatus.opLatencies
  • Labels: type

mongo_queued_operations​

  • Description: Number of operations queued waiting for a lock or a WiredTiger ticket
  • Type: Gauge
  • Source: serverStatus.globalLock.currentQueue
  • Labels: type

mongo_wt_tickets_available​

  • Description: Available WiredTiger concurrency tickets by type (read, write)
  • Type: Gauge
  • Source: serverStatus.queues.execution (8.0+) or serverStatus.wiredTiger.concurrentTransactions (7.0 and earlier)
  • Labels: type

mongo_wt_cache_used_bytes / mongo_wt_cache_dirty_bytes / mongo_wt_cache_max_bytes​

  • Description: WiredTiger cache usage, dirty bytes, and the configured maximum
  • Type: Gauge
  • Source: serverStatus.wiredTiger.cache

mongo_wt_pages_evicted_by_app_threads_total​

  • Description: Pages evicted from the WiredTiger cache by application threads - a key cache-pressure signal
  • Type: Counter
  • Source: serverStatus.wiredTiger.cache (sum of pages evicted by application threads, page evict attempts by application threads, and modified page evict attempts by application threads; the latter two are the 8.0+ counters)

mongo_wt_app_threads_evicting_seconds_total​

  • Description: Total time application threads spent evicting pages instead of serving queries
  • Type: Counter
  • Source: serverStatus.wiredTiger.cache (application thread time evicting (usecs))

mongo_wt_cache_bytes_read_into_total​

  • Description: Bytes read from disk into the WiredTiger cache (cache misses) - a direct signal that the working set does not fit in cache
  • Type: Counter
  • Source: serverStatus.wiredTiger.cache

mongo_wt_checkpoints_total / mongo_wt_checkpoint_seconds_total​

  • Description: Number of WiredTiger checkpoints completed and the cumulative time spent in them. The rate of the seconds counter is the fraction of wall-time spent checkpointing - sustained high values indicate write-stall risk.
  • Type: Counter
  • Source: serverStatus.wiredTiger (transaction on 7.0, checkpoint on 8.0+)

mongo_wt_journal_bytes_written_total​

  • Description: Bytes written to the WiredTiger journal (the crash-recovery write-ahead log)
  • Type: Counter
  • Source: serverStatus.wiredTiger.log

mongo_wt_journal_bytes_since_checkpoint​

  • Description: Bytes written to the journal since the last checkpoint - the amount that would be replayed on crash recovery (MongoDB's analog of Postgres's pg_wal_since_last_checkpoint_bytes)
  • Type: Gauge

mongo_time_since_last_checkpoint_seconds​

  • Description: Time since the last WiredTiger checkpoint observed by the agent; rises above the checkpoint interval if checkpoints stall
  • Type: Gauge

mongo_scanned_keys_total / mongo_scanned_documents_total​

  • Description: Total number of index keys / documents scanned by the query executor. Compared with documents returned to detect inefficient queries.
  • Type: Counter
  • Source: serverStatus.metrics.queryExecutor

mongo_collection_scans_total​

  • Description: Total number of queries that performed a collection scan
  • Type: Counter

mongo_scan_and_order_total​

  • Description: Total number of queries that performed an in-memory sort
  • Type: Counter

mongo_write_conflicts_total​

  • Description: Write conflicts (WiredTiger optimistic-concurrency retries); a rising rate indicates contention on hot documents
  • Type: Counter
  • Source: serverStatus.metrics.operation

mongo_ttl_deleted_documents_total​

  • Description: Total number of documents deleted by TTL indexes
  • Type: Counter

mongo_flow_control_time_acquiring_seconds_total​

  • Description: Total time write operations spent acquiring flow control tickets; a non-zero rate means flow control is throttling writes on the primary
  • Type: Counter
  • Source: serverStatus.flowControl

Per-query metrics​

Collected from the database profiler (system.profile), which requires operationProfiling to be enabled on the server, so the numbers cover slow (or, with Percona Server for MongoDB's rateLimit, sampled) operations. Query shapes are normalized (literal values replaced with ?); only the top 20 shapes by total execution time are reported. Rates are smoothed over a 60-second window.

Unlike the other MongoDB metrics (read from in-memory counters via serverStatus), these have a server-side cost: with profiling enabled, mongod writes a system.profile document for every captured operation. The agent only reads new entries incrementally, but the capture overhead is paid by the server — bound it with slowOpThresholdMs / mode: slowOp and, on Percona Server for MongoDB, rateLimit sampling. See Enabling the profiler for the trade-off.

mongo_top_query_calls_per_second / mongo_top_query_time_per_second / mongo_top_query_docs_returned_per_second / mongo_top_query_docs_examined_per_second / mongo_top_query_keys_examined_per_second​

  • Description: Per-query-shape execution rate, time, and documents/keys examined vs returned
  • Type: Gauge
  • Source: system.profile
  • Labels: db, collection, query

Current operations​

mongo_operations_waiting_for_lock​

  • Description: Number of operations waiting for a lock, by database
  • Type: Gauge
  • Source: $currentOp
  • Labels: db

mongo_long_running_operations​

  • Description: Number of operations of this query shape that have been running for at least 10s (their concurrency; top 20 shapes by count). Integrated over time this is the operation-time spent on long-running operations.
  • Type: Gauge
  • Source: $currentOp
  • Labels: db, collection, query, plan

mongo_fsync_locked​

  • Description: 1 if db.fsyncLock() is holding a global lock on this member (e.g. a filesystem-snapshot backup). On a secondary this blocks the oplog applier, so it is a direct root cause of replication lag.
  • Type: Gauge
  • Source: $currentOp (an operation with desc: "fsyncLockWorker" is present)

mongo_prepared_transactions​

  • Description: Number of transactions currently in the prepared state. A prepared transaction holds locks on secondaries until its commit/abort replicates, so a stuck or long one can block oplog apply.
  • Type: Gauge
  • Source: serverStatus.transactions (currentPrepared)

mongo_open_transactions​

  • Description: Number of open (including prepared) transactions broken down by the client application that owns them - a long or prepared one holds locks and can block oplog apply. The client name is normalized (UUID/pod suffixes stripped) and empty names are reported as unknown; only the top 20 applications are kept.
  • Type: Gauge
  • Source: $currentOp (entries with a transaction sub-document)
  • Labels: app

Cursors​

mongo_cursors_open / mongo_cursors_open_no_timeout / mongo_cursors_timed_out_total​

  • Description: Open cursors, open cursors created with the noTimeout option (a leak indicator), and the total number of cursors that timed out (a sign a client stopped iterating results before exhausting the cursor)
  • Type: Gauge (mongo_cursors_open, mongo_cursors_open_no_timeout), Counter (mongo_cursors_timed_out_total)
  • Source: serverStatus.metrics.cursor

Replication apply​

mongo_repl_apply_ops_total​

  • Description: Oplog operations applied by this member - the secondary apply throughput; a low rate on a lagging secondary indicates it is apply-bound
  • Type: Counter
  • Source: serverStatus.metrics.repl.apply

mongo_repl_buffer_operations / mongo_repl_buffer_bytes​

  • Description: Number and size of oplog operations buffered on a secondary waiting to be applied; a backed-up buffer explains replication lag
  • Type: Gauge
  • Source: serverStatus.metrics.repl.buffer

MongoDB backups​

When running on Kubernetes, backup state is collected through the agent's embedded kube-state-metrics from the custom resources of the Percona Operator for MongoDB (PerconaServerMongoDB, PerconaServerMongoDBBackup), which runs backups with Percona Backup for MongoDB (PBM). Every metric carries an operator label (percona) and, via the common namespace/name labels, correlates to the corresponding DatabaseCluster application in Coroot.

mongo_backup_target_info​

  • Description: A configured backup storage, one series per entry of spec.backup.storages (the method label is the storage name). Coroot assembles the destination from the S3 bucket and prefix or the Azure container.
  • Type: Info
  • Source: PerconaServerMongoDB.spec.backup.storages
  • Labels: operator, method, type, s3_bucket, s3_endpoint, s3_prefix, azure_container

mongo_backup_schedule_info​

  • Description: A scheduled backup task (one series per entry of spec.backup.tasks). Coroot uses the schedule of enabled tasks to detect overdue backups; method is the storage the task writes to and kind is logical or physical.
  • Type: Info
  • Source: PerconaServerMongoDB.spec.backup.tasks
  • Labels: operator, task, schedule, method, kind, enabled

mongo_backup_pitr_info​

  • Description: Whether point-in-time recovery (continuous oplog backup) is enabled for the cluster.
  • Type: Info
  • Source: PerconaServerMongoDB.spec.backup.pitr
  • Labels: operator, enabled

mongo_cluster_status​

  • Description: The cluster state reported by the operator (e.g. ready, initializing, error). Value is 1 for the currently-active series.
  • Type: Info
  • Source: PerconaServerMongoDB.status.state
  • Labels: operator, status

mongo_backup_info​

  • Description: An individual backup run (one series per PerconaServerMongoDBBackup object), used to list recent backups. Carries immutable identity only; the run's phase is in mongo_backup_status. cluster is the name of the PerconaServerMongoDB the backup belongs to.
  • Type: Info
  • Source: PerconaServerMongoDBBackup.spec (clusterName, storageName, type), .status.destination
  • Labels: operator, cluster, method, kind, path

mongo_backup_status​

  • Description: The current phase of a backup run (e.g. waiting, requested, running, ready, error). Value is 1 for the currently-active series (a run's phase changes over its lifecycle, so only the live series is used).
  • Type: Info
  • Source: PerconaServerMongoDBBackup.status.state
  • Labels: operator, status

mongo_backup_completed_timestamp_seconds​

  • Description: When a backup run completed.
  • Type: Gauge
  • Source: PerconaServerMongoDBBackup.status.completed
  • Labels: operator

Change tracking​

When --track-database-changes is enabled, the agent detects and emits change events for:

  • Index changes — MongoDB is schemaless, so instead of DDL the agent tracks indexes per collection. Each cycle it snapshots all indexes (name and key fields) for every collection. When indexes are added, removed, or modified, a change event is emitted with a unified diff.
  • Settings changes — The agent snapshots all server parameters via getParameter: "*" each cycle. When a parameter changes (e.g., after setParameter or a restart), a change event is emitted with the diff.

MongoDB system databases (admin, config, local) are always excluded from index tracking.

Each change event includes db.system, db.target, db.name, db_change.object, and db_change.type attributes.

Size metrics​

The agent collects database and collection size metrics. For collection sizes, only the top 20 largest collections across all databases are reported. MongoDB system databases (admin, config, local) are always excluded.

mongo_database_size_bytes​

  • Description: Total size of the database in bytes
  • Type: Gauge
  • Source: listDatabases command (sizeOnDisk)
  • Labels: db

mongo_collection_size_bytes​

  • Description: Total size of the collection in bytes (data + indexes + storage overhead)
  • Type: Gauge
  • Source: collStats command (totalSize)
  • Labels: db, collection

mongo_collection_size_growth_bytes_per_second​

  • Description: Collection size growth rate in bytes per second. Only the top 20 fastest growing collections across all databases are reported. Requires at least two collection cycles to compute.
  • Type: Gauge
  • Source: Computed from consecutive $collStats measurements
  • Labels: db, collection

mongo_collection_storage_size_bytes​

  • Description: Bytes allocated on disk for documents of the collection
  • Type: Gauge
  • Source: $collStats (storageStats.storageSize)
  • Labels: db, collection

mongo_collection_free_storage_bytes​

  • Description: Reusable (fragmented) bytes within the allocated collection storage
  • Type: Gauge
  • Source: $collStats (storageStats.freeStorageSize)
  • Labels: db, collection

mongo_collection_documents​

  • Description: Number of documents in the collection
  • Type: Gauge
  • Source: $collStats (storageStats.count)
  • Labels: db, collection

Redis​

Redis metrics are collected by the embedded redis_exporter (in redis-metrics-only mode with latency histograms disabled), so the full metric set and its semantics are described in the exporter's documentation. The metrics Coroot relies on are:

redis_up​

  • Description: Whether the Redis server is reachable or not
  • Type: Gauge

redis_exporter_last_scrape_error​

  • Description: Whether a scrape error occurred
  • Type: Gauge
  • Labels: err

redis_instance_info​

  • Description: The server info; Coroot uses role (master/slave) to build the replication topology
  • Type: Gauge
  • Labels: redis_version, role, and other fields of INFO server/INFO replication

redis_commands_total / redis_commands_duration_seconds_total​

  • Description: Total number of calls and cumulative execution time per command. The rate of the seconds counter divided by the rate of calls is the average command latency.
  • Type: Counter
  • Source: INFO commandstats
  • Labels: cmd

redis_db_keys / redis_db_keys_expiring​

  • Description: Number of keys and number of keys with a TTL in each logical database
  • Type: Gauge
  • Source: INFO keyspace
  • Labels: db

Memcached​

Memcached metrics are collected by the embedded memcached_exporter, so the full metric set is described in the exporter's documentation. The metrics Coroot relies on are:

memcached_up​

  • Description: Whether the Memcached server is reachable or not
  • Type: Gauge

memcached_version​

  • Description: The server version
  • Type: Gauge
  • Labels: version

memcached_limit_bytes​

  • Description: The configured memory limit for item storage (-m)
  • Type: Gauge

memcached_items_evicted_total​

  • Description: Total number of valid items removed from the cache to free memory for new items; a non-zero rate means the cache is undersized for the working set
  • Type: Counter

memcached_commands_total​

  • Description: Total number of commands by type and outcome (get/hit, get/miss, set, delete, ...); Coroot derives the hit rate from the get hits and misses
  • Type: Counter
  • Labels: command, status

AWS​

When the AWS integration is configured, the agent discovers RDS instances and ElastiCache nodes through the AWS API (optionally filtered by tags) and exposes their state. Every RDS metric carries an rds_instance_id label (<region>/<DBInstanceIdentifier>) and every ElastiCache metric an ec_instance_id label (<region>/<CacheClusterId>/<CacheNodeId>), which Coroot uses to match the instances to the applications that connect to them.

aws_discovery_error​

  • Description: 1 for each distinct AWS API error encountered during the last discovery cycle, 0 when discovery succeeded
  • Type: Gauge
  • Labels: error

aws_rds_info​

  • Description: RDS instance info
  • Type: Gauge
  • Source: DescribeDBInstances; ipv4 is resolved by the agent from the endpoint address
  • Labels: region, availability_zone, endpoint, ipv4, port, engine, engine_version, instance_type, storage_type, multi_az, secondary_availability_zone, cluster_id, source_instance_id

aws_rds_status​

  • Description: The status of the RDS instance (e.g. available, modifying, backing-up)
  • Type: Gauge
  • Labels: status

aws_rds_allocated_storage_gibibytes / aws_rds_storage_autoscaling_threshold_gibibytes / aws_rds_storage_provisioned_iops​

  • Description: The allocated storage size, the storage autoscaling upper limit (MaxAllocatedStorage), and the number of provisioned IOPS
  • Type: Gauge

aws_rds_backup_retention_period_days​

  • Description: The number of days automated backups are retained
  • Type: Gauge

aws_rds_read_replica_info​

  • Description: One series per read replica of this instance
  • Type: Gauge
  • Labels: replica_instance_id

aws_rds_log_messages_total​

  • Description: Number of messages in the instance's logs (postgres, aurora-postgresql, mysql, mariadb and aurora-mysql engines) grouped by the automatically extracted repeated pattern
  • Type: Counter
  • Source: the instance's log files, read through the RDS DownloadDBLogFilePortion API
  • Labels: level, pattern_hash, sample

The following OS-level metrics are read from RDS Enhanced Monitoring (the RDSOSMetrics CloudWatch Logs group) and are only available when Enhanced Monitoring is enabled for the instance:

aws_rds_cpu_cores​

  • Description: The number of virtual CPUs
  • Type: Gauge

aws_rds_cpu_usage_percent​

  • Description: The percentage of the CPU spent in each mode
  • Type: Gauge
  • Labels: mode (user, system, wait, steal, irq, nice, guest)

aws_rds_memory_total_bytes / aws_rds_memory_cached_bytes / aws_rds_memory_free_bytes​

  • Description: The total amount of memory, the amount used as page cache, and the amount of unassigned memory
  • Type: Gauge

aws_rds_io_ops_per_second / aws_rds_io_bytes_per_second​

  • Description: The number of I/O operations and bytes read or written per second, per device (aurora-data for Aurora's network storage)
  • Type: Gauge
  • Labels: device, operation (read, write)

aws_rds_io_await_seconds / aws_rds_io_util_percent​

  • Description: The average time to serve an I/O request including queue time, and the percentage of time during which requests were issued to the device
  • Type: Gauge
  • Labels: device

aws_rds_io_latency_seconds​

  • Description: The average elapsed time between the submission of an I/O request and its completion (Amazon Aurora only)
  • Type: Gauge
  • Labels: device, operation

aws_rds_fs_total_bytes / aws_rds_fs_used_bytes​

  • Description: The size of each file system and the space used by files on it; Coroot uses the /rdsdbdata mount point for the data volume
  • Type: Gauge
  • Labels: mount_point

aws_rds_net_rx_bytes_per_second / aws_rds_net_tx_bytes_per_second​

  • Description: The number of bytes received and transmitted per second, per network interface
  • Type: Gauge
  • Labels: interface

aws_elasticache_info​

  • Description: ElastiCache node info; cluster_id is the replication group id when the node belongs to one, the cache cluster id otherwise
  • Type: Gauge
  • Source: DescribeCacheClusters; ipv4 is resolved by the agent from the endpoint address
  • Labels: region, availability_zone, endpoint, ipv4, port, engine, engine_version, instance_type, cluster_id

aws_elasticache_status​

  • Description: The status of the ElastiCache node (e.g. available, creating, rebooting)
  • Type: Gauge
  • Labels: status

FluxCD​

The agent's embedded kube-state-metrics reads FluxCD custom resources (source.toolkit.fluxcd.io, kustomize.toolkit.fluxcd.io, helm.toolkit.fluxcd.io, fluxcd.controlplane.io) and exposes their state. This requires the agent's service account to have get/list/watch access to those API groups, which the Coroot Operator grants automatically.

Every metric below also carries uid, name, and namespace labels identifying the source custom resource. The *_info metrics are info-style: their value is always 1 and the useful data is carried in labels. The *_status metrics expose Kubernetes status conditions: one series per condition, with the value 1 when the condition holds (status: "True") and 0 otherwise ("False"/"Unknown").

fluxcd_git_repository_info / fluxcd_oci_repository_info / fluxcd_helm_repository_info​

  • Description: Information about a GitRepository / OCIRepository / HelmRepository source
  • Type: Info
  • Source: the source spec
  • Labels: url, interval, suspended

fluxcd_git_repository_status / fluxcd_oci_repository_status / fluxcd_helm_repository_status​

  • Description: Status conditions of a GitRepository / OCIRepository / HelmRepository source
  • Type: Gauge
  • Source: status.conditions[]
  • Labels: type (condition type, e.g. Ready), reason

fluxcd_helm_release_info​

  • Description: Information about a HelmRelease
  • Type: Info
  • Source: the HelmRelease spec
  • Labels: suspended, interval, target_namespace, source_kind, source_name, source_namespace, chart, version, chart_ref_kind, chart_ref_name, chart_ref_namespace

fluxcd_helm_release_status​

  • Description: Status conditions of a HelmRelease
  • Type: Gauge
  • Source: status.conditions[]
  • Labels: type, reason

fluxcd_helm_chart_info​

  • Description: Information about a HelmChart
  • Type: Info
  • Source: the HelmChart spec
  • Labels: chart, version, source_kind, source_name, source_namespace, interval, suspended

fluxcd_helm_chart_status​

  • Description: Status conditions of a HelmChart
  • Type: Gauge
  • Source: status.conditions[]
  • Labels: type, reason

fluxcd_kustomization_info​

  • Description: Information about a Kustomization
  • Type: Info
  • Source: the Kustomization spec and status
  • Labels: suspended, interval, path, source_kind, source_name, source_namespace, target_namespace, last_applied_revision, last_attempted_revision

fluxcd_kustomization_status​

  • Description: Status conditions of a Kustomization
  • Type: Gauge
  • Source: status.conditions[]
  • Labels: type, reason

fluxcd_kustomization_inventory_entry_info​

  • Description: A resource managed by a Kustomization (one series per inventory entry)
  • Type: Info
  • Source: status.inventory.entries[]
  • Labels: entry_id

fluxcd_kustomization_dependency_info​

  • Description: A dependency declared by a Kustomization (one series per dependsOn entry)
  • Type: Info
  • Source: spec.dependsOn[]
  • Labels: depends_on_name, depends_on_namespace

fluxcd_resourceset_info​

  • Description: Information about a ResourceSet
  • Type: Info
  • Source: the ResourceSet status
  • Labels: last_applied_revision

fluxcd_resourceset_status​

  • Description: Status conditions of a ResourceSet
  • Type: Gauge
  • Source: status.conditions[]
  • Labels: type, reason

fluxcd_resourceset_inventory_entry_info​

  • Description: A resource managed by a ResourceSet (one series per inventory entry)
  • Type: Info
  • Source: status.inventory.entries[]
  • Labels: entry_id

fluxcd_resourceset_dependency_info​

  • Description: A dependency declared by a ResourceSet (one series per dependsOn entry)
  • Type: Info
  • Source: spec.dependsOn[]
  • Labels: depends_on_kind, depends_on_name, depends_on_namespace

ArgoCD​

The agent's embedded kube-state-metrics reads ArgoCD Application resources (argoproj.io/v1alpha1) and exposes their sync, health, and operation state. This requires the agent's service account to have get/list/watch access to the argoproj.io API group, which the Coroot Operator grants automatically.

Every metric below also carries uid, name, and namespace labels identifying the Application. All of these are info-style metrics: their value is always 1 and the meaningful state is carried in labels (such as sync_status), so a status that ArgoCD doesn't currently report simply has no series.

argocd_application_info​

  • Description: Information about an Application
  • Type: Info
  • Source: the Application spec and status
  • Labels: project, source_type, repo, path, chart, target_revision, dest_server, dest_name, dest_namespace, revision

argocd_application_sync_status​

  • Description: Sync status of an Application
  • Type: Info
  • Source: status.sync.status
  • Labels: sync_status (e.g. Synced, OutOfSync, Unknown)

argocd_application_health_status​

  • Description: Health status of an Application
  • Type: Info
  • Source: status.health.status
  • Labels: health_status (e.g. Healthy, Progressing, Degraded, Suspended, Missing, Unknown)

argocd_application_operation_status​

  • Description: Phase of the most recent sync operation
  • Type: Info
  • Source: status.operationState.phase
  • Labels: operation_phase (e.g. Running, Succeeded, Failed, Error, Terminating)

argocd_application_operation_finished_timestamp_seconds​

  • Description: When the most recent sync operation finished, as a Unix timestamp
  • Type: Gauge
  • Source: status.operationState.finishedAt

argocd_application_resource_info​

  • Description: A resource managed by an Application (one series per resource)
  • Type: Info
  • Source: status.resources[]
  • Labels: resource_group, resource_kind, resource_namespace, resource_name

argocd_application_resource_sync_status​

  • Description: Sync status of a managed resource
  • Type: Info
  • Source: status.resources[].status
  • Labels: resource_group, resource_kind, resource_namespace, resource_name, sync_status

argocd_application_resource_health_status​

  • Description: Health status of a managed resource
  • Type: Info
  • Source: status.resources[].health.status
  • Labels: resource_group, resource_kind, resource_namespace, resource_name, health_status

argocd_application_resource_status​

  • Description: Result for a resource from the most recent sync operation
  • Type: Info
  • Source: status.operationState.syncResult.resources[]
  • Labels: resource_group, resource_kind, resource_namespace, resource_name, status (e.g. Synced, Pruned, SyncFailed)

GCP​

When the GCP integration is configured, the agent discovers Cloud SQL and Memorystore instances through the GCP APIs (optionally filtered by labels) and exposes their state. Every Cloud SQL metric carries a cloudsql_instance_id label (<project>/<instance>) and every Memorystore metric a memorystore_instance_id label (<project>/<region>/<instance>), which Coroot uses to match the instances to the applications that connect to them.

gcp_discovery_error​

  • Description: 1 for each distinct GCP API error encountered during the last discovery cycle, 0 when discovery succeeded
  • Type: Gauge
  • Labels: error

gcp_cloudsql_info​

  • Description: Cloud SQL instance info
  • Type: Gauge
  • Source: the Cloud SQL Admin API (instances.list); ipv4 is the private IP if the instance has one, otherwise the public one
  • Labels: project, region, zone, ipv4, port, engine, engine_version, tier, availability_type, connection_name, instance_type (CLOUD_SQL_INSTANCE or READ_REPLICA_INSTANCE), primary_instance (the primary of a read replica)

gcp_cloudsql_status​

  • Description: The state of the Cloud SQL instance (e.g. RUNNABLE, SUSPENDED, MAINTENANCE)
  • Type: Gauge
  • Labels: status

gcp_cloudsql_cpu_usage_percent / gcp_cloudsql_cpu_cores / gcp_cloudsql_cpu_usage_cores​

  • Description: CPU utilization as a percentage of the reserved vCPUs (database/cpu/utilization), the number of reserved vCPUs (database/cpu/reserved_cores), and the CPU time of the database process in cores (database/cpu/usage_time aligned as a rate)
  • Type: Gauge
  • Source: Cloud Monitoring (cloudsql.googleapis.com/database/*), the latest 1-minute aligned value

gcp_cloudsql_memory_total_bytes / gcp_cloudsql_memory_used_bytes / gcp_cloudsql_memory_components_percent​

  • Description: the memory quota (database/memory/quota), the memory usage of the database process including its buffers and cache (database/memory/total_usage), and the quota split into the usage, cache and free components in percent (database/memory/components)
  • Type: Gauge
  • Source: Cloud Monitoring, the latest 1-minute aligned value
  • Labels: component (usage, cache, free) for gcp_cloudsql_memory_components_percent

gcp_cloudsql_disk_total_bytes / gcp_cloudsql_disk_used_bytes​

  • Description: the data disk quota (database/disk/quota) and usage (database/disk/bytes_used)
  • Type: Gauge
  • Source: Cloud Monitoring, the latest 1-minute aligned value

gcp_cloudsql_io_ops_per_second​

  • Description: Disk I/O operations per second of the instance
  • Type: Gauge
  • Source: Cloud Monitoring, disk/read_ops_count and disk/write_ops_count aligned as rates
  • Labels: operation (read, write)

gcp_cloudsql_io_bytes_per_second​

  • Description: Disk I/O throughput of the instance
  • Type: Gauge
  • Source: Cloud Monitoring, disk/read_bytes_count and disk/write_bytes_count aligned as rates
  • Labels: operation (read, write)

gcp_cloudsql_network_bytes_per_second​

  • Description: Network throughput of the instance
  • Type: Gauge
  • Source: Cloud Monitoring, network/received_bytes_count and network/sent_bytes_count aligned as rates
  • Labels: direction (rx, tx)

gcp_cloudsql_log_messages_total​

  • Description: Number of messages in the instance's logs (postgres and mysql engines) grouped by the automatically extracted repeated pattern
  • Type: Counter
  • Source: Cloud Logging, the cloudsql_database resource of the instance
  • Labels: level, pattern_hash, sample

gcp_memorystore_info​

  • Description: Memorystore instance info. Redis and Valkey instances have one series, Memcached instances one per node (memorystore_instance_id is <project>/<region>/<instance>/<node>)
  • Type: Gauge
  • Source: the Memorystore for Redis, Memcached and Valkey APIs (instances.list)
  • Labels: project, region, zone, ipv4, port, engine (redis, memcached, valkey), engine_version, tier, memory_size_gb, instance

gcp_memorystore_status​

  • Description: The state of the Memorystore instance (READY for Redis and Memcached, ACTIVE for Valkey, or a transitional state)
  • Type: Gauge
  • Labels: status

gcp_memorystore_cpu_usage_percent / gcp_memorystore_cpu_usage_cores / gcp_memorystore_memory_used_bytes / gcp_memorystore_network_bytes_per_second​

  • Description: OS-level metrics of the node from Cloud Monitoring, what each product publishes: CPU utilization in percent (Valkey: instance/cpu/average_utilization of the primaries), CPU time of the engine process in cores (Redis: stats/cpu_utilization of the primary summed over user and system time, Memcached: node/cpu/usage_time summed over the modes), memory used by the engine (Redis: stats/memory/usage, Valkey: instance/memory/total_used_memory, Memcached: the used part of node/cache_memory) and network throughput (Redis: stats/network_traffic, Memcached: node/received_bytes_count and node/sent_bytes_count, aligned as rates)
  • Type: Gauge
  • Source: Cloud Monitoring (redis.googleapis.com/, memorystore.googleapis.com/instance/, memcache.googleapis.com/node/), the latest 1-minute aligned value
  • Labels: direction (rx, tx) for gcp_memorystore_network_bytes_per_second

gcp_memorystore_cpu_cores / gcp_memorystore_memory_total_bytes​

  • Description: The vCPU count and memory capacity of the node. Memory comes from the instance configuration of all three products; vCPUs are reported only for Memcached nodes (nodeConfig.cpuCount), the Redis and Valkey APIs don't return them
  • Type: Gauge

OCI​

When the OCI integration is configured, the agent discovers the MySQL HeatWave and Database with PostgreSQL DB systems and the OCI Cache clusters of the configured compartments. Every DB system metric carries an oci_db_id label and every cache metric an oci_cache_id label: the OCID of the resource, or the instance id for the standby instances of a PostgreSQL DB system.

oci_discovery_error​

  • Description: 1 for each distinct OCI API error encountered since the previous discovery cycle (discovery, OCI Monitoring and the log readers), 0 when there was none
  • Type: Gauge
  • Labels: error

oci_db_info​

  • Description: DB system info
  • Type: Gauge
  • Source: the MySQL HeatWave (ListDbSystems, ListReplicas) and Database with PostgreSQL (ListDbSystems, GetDbSystem, GetPrimaryDbInstance, GetConnectionDetails) APIs
  • Labels: name, compartment, region, availability_domain (in the node-agent's form, e.g. us-ashburn-1-ad-1), ipv4, port, engine (mysql, postgres), engine_version, shape, high_availability, primary (the name of the primary for MySQL read replicas and PostgreSQL standby instances, which are reported as separate instances)

oci_db_status​

  • Description: The lifecycle state of the DB system (e.g. ACTIVE, UPDATING, INACTIVE)
  • Type: Gauge
  • Labels: status

oci_db_cpu_cores / oci_db_memory_total_bytes​

  • Description: the vCPUs (2 per OCPU) and the memory of the DB system, from the shape (PostgreSQL) or from OCI Monitoring (OCPUsAllocated, MemoryAllocated for MySQL)
  • Type: Gauge

oci_db_cpu_usage_percent / oci_db_cpu_usage_cores / oci_db_memory_used_bytes / oci_db_memory_usage_percent​

  • Description: CPU utilization (CPUUtilization), CPU usage in vCPUs (MySQL: OCPUsUsed), memory used by the database (MySQL: MemoryUsed) and memory utilization (PostgreSQL: MemoryUtilization)
  • Type: Gauge
  • Source: OCI Monitoring (oci_mysql_database, oci_postgresql), the latest 1-minute value

oci_db_disk_total_bytes / oci_db_disk_used_bytes​

  • Description: storage allocated and used (MySQL: StorageAllocated, StorageUsed; PostgreSQL: UsedStorage)
  • Type: Gauge

oci_db_io_ops_per_second / oci_db_io_bytes_per_second / oci_db_io_latency_seconds / oci_db_network_bytes_per_second​

  • Description: disk I/O operations and throughput (DbVolumeRead/WriteOperations, DbVolumeRead/WriteBytes, Read/WriteIops, Read/WriteThroughput) and network throughput (NetworkReceive/TransmitBytes, MySQL only)
  • Type: Gauge
  • Labels: operation (read, write), direction (rx, tx)
  • Note: the I/O latency (ReadLatency, WriteLatency) is reported for PostgreSQL only

oci_db_log_messages_total​

  • Description: the number of messages in the DB system's log grouped by the automatically extracted repeated pattern
  • Labels: level, pattern_hash, sample
  • Source: OCI Logging, the postgresql_database_logs service log of the DB system (PostgreSQL), or the server's error log read through performance_schema.error_log (MySQL HeatWave, which publishes no logs to OCI Logging)

oci_cache_info​

  • Description: OCI Cache cluster info
  • Type: Gauge
  • Source: the OCI Cache API (ListRedisClusters); ipv4 is the primary endpoint
  • Labels: name, compartment, region, ipv4, port, engine (valkey, redis), engine_version, node_count, node_memory_gb

oci_cache_status​

  • Description: The lifecycle state of the cluster (ACTIVE or a transitional state)
  • Type: Gauge
  • Labels: status

oci_cache_memory_total_bytes / oci_cache_cpu_usage_percent / oci_cache_memory_used_bytes / oci_cache_network_bytes_per_second​

  • Description: the memory of a node, and from OCI Monitoring (oci_redis): CPU utilization (CPUUtilization), memory used by the engine (UsedMemory) and network throughput (NetworkBytesIn/Out)
  • Type: Gauge
  • Labels: direction (rx, tx) for the network metric

oci_cache_log_messages_total​

  • Description: the number of messages in the cache cluster's engine log grouped by the automatically extracted repeated pattern
  • Labels: level, pattern_hash, sample
  • Source: OCI Logging, the oci-cache-engine-logs service log of the cluster
Looking for 24/7 support from the Coroot team? Subscribe to Coroot Enterprise:Start free trial