This topic describes the native monitoring metrics provided by OB Cloud Exporter, including common labels, raw units, and value semantics.
The list covers the metrics currently available, but not every request returns all metrics. The metrics returned depend on factors such as the OceanBase version, deployment type, whether the tenant or cluster exists, and whether collection succeeds.
The units and metric definitions in this topic apply to the native metrics returned by Exporter. Native metrics may share names with logical metrics that the console derives through calculations or unit conversions. Do not use their units or alert thresholds interchangeably. For integration instructions, see Monitor OceanBase Cloud with Prometheus.
Conventions
- Value semantics: A
counterrepresents a cumulative value and is typically used withrate()orincrease(). Agaugerepresents a current value and is typically read directly. Process restarts or statistics resets may reset cumulative values to zero. - Units: Byte denotes bytes. KiB, MiB, and GiB use base 1024. The units s, ms, μs, and ns denote seconds, milliseconds, microseconds, and nanoseconds, respectively. Gbit/s denotes 10⁹ bit/s.
Common labels
The following table describes common labels.
Label |
Description |
|---|---|
app |
Application type (HOST: host; OB: OceanBase) |
svr_ip, host_name |
Node IP address and cloud host identifier |
ob_cluster_name |
Cluster ID or identifier |
obregion, obzone |
OceanBase cluster name and OceanBase zone name |
ob_tenant_id, tenant_name |
Tenant ID and tenant name |
uid |
Cloud account identifier |
obproxy_cluster |
OBProxy cluster ID |
ob_cluster_name_alias, tenant_name_alias |
Cluster display alias and tenant display alias |
device, dev |
Device name |
mountpoint, fstype |
File system mount point and file system type |
mount_label, ob_device, is_ob_disk, is_ob_volume |
Disk purpose labels that distinguish installation disks, data disks, log disks, and other disk types |
name |
Process name |
stat_id |
Identifier of an ob_sysstat statistic |
Host and OceanBase metrics
Basic host metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
node_cpu_seconds_total |
Cumulative CPU time aggregated by mode |
s | counter |
node_filesystem_avail_bytes |
Free file system space available to unprivileged users | Byte | gauge |
node_filesystem_size_bytes |
Total file system capacity | Byte | gauge |
node_disk_reads_completed_total |
Cumulative number of completed disk read operations | count | counter |
node_disk_read_bytes_total |
Cumulative bytes read from disk | Byte | counter |
node_disk_read_time_seconds_total |
Cumulative time spent on disk read operations | s | counter |
node_disk_writes_completed_total |
Cumulative number of completed disk write operations | count | counter |
node_disk_written_bytes_total |
Cumulative bytes written to disk | Byte | counter |
node_disk_write_time_seconds_total |
Cumulative time spent on disk write operations | s | counter |
node_load1 |
System load average over the last 1 minute | dimensionless | gauge |
node_load5 |
System load average over the last 5 minutes | dimensionless | gauge |
node_load15 |
System load average over the last 15 minutes | dimensionless | gauge |
node_memory_Buffers_bytes |
Kernel buffer memory | Byte | gauge |
node_memory_MemFree_bytes |
Free memory | Byte | gauge |
node_memory_Cached_bytes |
Page cache memory | Byte | gauge |
node_memory_MemTotal_bytes |
Total memory | Byte | gauge |
node_network_receive_bytes_total |
Cumulative bytes received by the network interface | Byte | counter |
node_network_transmit_bytes_total |
Cumulative bytes transmitted by the network interface | Byte | counter |
node_boot_time_seconds |
Unix timestamp of system startup | s | gauge |
node_time_seconds |
System Unix timestamp at collection time | s | gauge |
Custom host metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
io_util |
Percentage of the sampling period spent processing disk I/O | % | gauge |
io_await |
Average time per disk read or write operation during the sampling period | ms | gauge |
node_ntp_offset_seconds |
Clock offset reported by NTP or Chrony | s | gauge |
node_ntp_system_time_offset_seconds |
Offset between system time and NTP time reported by Chrony | s | gauge |
ntp_service_available |
Whether NTP offset collection succeeded: 1 for success, 0 for command execution or parsing failure | dimensionless | gauge |
ntp_suitable_source_available |
Reports 0 when chronyd -Q detects no suitable time source (currently reports only this error condition) |
dimensionless | gauge |
Process metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
process_exists |
Whether the specified process exists: 1 for yes, 0 for no | dimensionless | gauge |
process_count |
Number of observer processes with the same name | count | gauge |
process_uptime_seconds |
Uptime of the observer process | s | gauge |
process_boot_time_seconds |
Unix timestamp of observer process startup | s | gauge |
process_memory_rss_bytes |
Resident set size (RSS) of the observer process | Byte | gauge |
process_cpu_percent |
CPU utilization of the observer process | % | gauge |
process_fd_count |
Number of file descriptors for the observer process | count | gauge |
process_thread_count |
Number of threads in the observer process | count | gauge |
OceanBase SQL metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
ob_role_is_rootserver |
Whether the current OBServer hosts RootService: 1 for yes, 0 for no | dimensionless | gauge |
ob_sysstat |
Statistical value from v$sysstat; the meaning and unit depend on stat_id |
See the stat_id tables |
mixed |
ob_cache_size_bytes |
Current memory usage of the tenant's KVCache types | Byte | gauge |
ob_disk_total_bytes |
Data disk capacity of the current OBServer | Byte | gauge |
ob_disk_free_bytes |
Free data disk space on the current OBServer | Byte | gauge |
ob_plan_cache_memory_bytes |
Memory used by the tenant's plan cache | Byte | gauge |
ob_plan_cache_access_total |
Cumulative accesses to the tenant's plan cache | count | counter |
ob_plan_cache_hit_total |
Cumulative hits in the tenant's plan cache | count | counter |
ob_active_session_num |
Current number of ACTIVE sessions for the tenant | count | gauge |
ob_all_session_num |
Current total number of sessions for the tenant | count | gauge |
ob_server_uptime_seconds |
Uptime of the current node from start_service_time to collection time |
s | gauge |
ob_query_rt_cumulative_count |
Cumulative number of requests with SQL response times less than or equal to le |
count; le is in μs |
cumulative bucket count |
ob_system_event_total_waits |
Cumulative non-IDLE waits aggregated by wait class | count | counter |
ob_system_event_time_waited |
Cumulative non-IDLE wait time aggregated by wait class | s | counter |
ob_tenant_available_disk_total |
Largest total OceanBase data disk capacity among the zones hosting the tenant | Byte | gauge |
ob_tenant_log_disk_total_bytes |
Log disk capacity allocated to the tenant on the current node | Byte | gauge |
ob_tenant_log_disk_used_bytes |
Log disk space used by the tenant on the current node | Byte | gauge |
ob_tenant_server_required_size |
Disk space actually occupied by the tenant's data files on the node | Byte | gauge |
ob_tenant_server_data_size |
Size of the tenant's data on the node | Byte | gauge |
ob_waitevent_wait_total |
Cumulative number of all non-IDLE waits for the tenant | count | counter |
ob_waitevent_wait_seconds_total |
Cumulative duration of all non-IDLE waits for the tenant | s | counter |
ob_foreground_waitevent_cnt |
Number of foreground non-IDLE wait or ON_CPU samples in ASH over the last second | count | gauge |
ob_background_waitevent_cnt |
Number of background non-IDLE wait or ON_CPU samples in ASH over the last second | count | gauge |
ob_sql_event_in_parse |
Number of ASH samples in the SQL parsing stage over the last second | count | gauge |
ob_sql_event_in_pl_parse |
Number of ASH samples in the PL parsing stage over the last second | count | gauge |
ob_sql_event_in_plan_cache |
Number of ASH samples in the plan cache access stage over the last second | count | gauge |
ob_sql_event_in_sql_optimize |
Number of ASH samples in the SQL optimization stage over the last second | count | gauge |
ob_sql_event_in_sql_execution |
Number of ASH samples in the SQL execution stage over the last second | count | gauge |
ob_sql_event_in_px_execution |
Number of ASH samples in the PX parallel execution stage over the last second | count | gauge |
ob_sql_event_in_sequence_load |
Number of ASH samples retrieving auto-increment column or SEQUENCE values over the last second | count | gauge |
ob_ss_cache_total_hit_cnt |
Cumulative local cache hits in Shared Storage | count | counter |
ob_ss_cache_total_miss_cnt |
Cumulative local cache misses in Shared Storage | count | counter |
ob_tenant_space_usage_bytes |
Tenant space usage by SPACE_TYPE |
Byte | gauge |
ob_tenant_vector_mem_hold_bytes |
Memory held by vector indexes | Byte | gauge |
ob_tenant_vector_mem_used_bytes |
Memory used by vector indexes | Byte | gauge |
ob_tenant_vector_mem_limit_bytes |
Memory usage limit for vector indexes | Byte | gauge |
ob_clog_ls_end_scn |
End SCN of contiguous logs in the log stream | SCN, in nanoseconds | gauge |
ob_clog_ls_max_replayed_scn |
Highest SCN replayed in the log stream | SCN, in nanoseconds | gauge |
ob_m_disk_tenant_base_data_size |
Size of the tenant's baseline data in major SSTables, with duplicate replicas excluded according to upstream rules | Byte | gauge |
ob_user_disk_total_bytes |
Node data disk capacity configured by the customer; the actual unit is GiB despite the _bytes suffix |
GiB | gauge |
ob_m_ssd_threshold_iops |
Maximum IOPS of an individual SSD or cloud disk | count/s | gauge |
ob_m_ssd_threshold_throughput |
Maximum throughput of an individual SSD or cloud disk | MiB/s | gauge |
ob_m_ecs_threshold_iops |
ECS instance-level disk IOPS limit | count/s | gauge |
ob_m_ecs_threshold_throughput |
ECS instance-level disk throughput limit | Gbit/s | gauge |
ob_sysstat statistics
Example metric with labels (some labels omitted): ob_sysstat{stat_id="40000",ob_tenant_id="…",tenant_name="…",…}
Network and request queues
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 10000 | Cumulative number of RPC packets received | count | counter |
| 10001 | Cumulative size of RPC packets received | Byte | counter |
| 10002 | Cumulative number of RPC packets sent | count | counter |
| 10003 | Cumulative size of RPC packets sent | Byte | counter |
| 10005 | Cumulative RPC network latency | μs | counter |
| 10006 | Cumulative RPC framework latency | μs | counter |
| 20000 | Cumulative number of requests enqueued | count | counter |
| 20001 | Cumulative number of requests dequeued | count | counter |
| 20002 | Cumulative time spent waiting in the request queue | μs | counter |
Transactions and clog synchronization
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 30000 | Cumulative clog synchronization time | μs | counter |
| 30001 | Cumulative number of clog synchronizations | count | counter |
| 30002 | Cumulative number of clog submissions | count | counter |
| 30006 | Total transaction time | μs | counter |
| 30007 | Cumulative number of transaction commits | count | counter |
| 30008 | Cumulative time spent committing transactions | μs | counter |
| 30009 | Cumulative number of transaction rollbacks | count | counter |
| 30010 | Cumulative time spent rolling back transactions | μs | counter |
| 30011 | Cumulative number of transaction timeouts | count | counter |
| 30012 | Cumulative number of local transactions | count | counter |
| 30013 | Cumulative number of distributed transactions | count | counter |
SQL and cursors
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 40000 | Cumulative number of SELECT statements executed | count | counter |
| 40001 | Cumulative SELECT statement execution time | μs | counter |
| 40002 | Cumulative number of INSERT statements executed | count | counter |
| 40003 | Cumulative INSERT statement execution time | μs | counter |
| 40004 | Cumulative number of REPLACE statements executed | count | counter |
| 40005 | Cumulative REPLACE statement execution time | μs | counter |
| 40006 | Cumulative number of UPDATE statements executed | count | counter |
| 40007 | Cumulative UPDATE statement execution time | μs | counter |
| 40008 | Cumulative number of DELETE statements executed | count | counter |
| 40009 | Cumulative DELETE statement execution time | μs | counter |
| 40010 | Cumulative number of local SQL executions | count | counter |
| 40011 | Cumulative number of remote SQL executions | count | counter |
| 40012 | Cumulative number of distributed SQL executions | count | counter |
| 40018 | Cumulative number of other SQL statements executed | count | counter |
| 40019 | Cumulative execution time of other SQL statements | μs | counter |
| 40030 | Current number of open cursors (opened cursors current) | count | gauge |
Cache hits and I/O
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 50000 | Cumulative row cache hits | count | counter |
| 50001 | Cumulative row cache misses | count | counter |
| 50004 | Cumulative Bloom filter cache hits | count | counter |
| 50005 | Cumulative Bloom filter cache misses | count | counter |
| 50008 | Cumulative block cache hits | count | counter |
| 50009 | Cumulative block cache misses | count | counter |
| 50010 | Cumulative location cache hits | count | counter |
| 50011 | Cumulative location cache misses | count | counter |
| 50037 | Cumulative clog cache hits | count | counter |
| 50038 | Cumulative clog cache misses | count | counter |
| 60000 | Cumulative number of I/O reads | count | counter |
| 60001 | Cumulative I/O read latency | μs | counter |
| 60002 | Cumulative bytes read by I/O operations | Byte | counter |
| 60003 | Cumulative number of I/O writes | count | counter |
| 60004 | Cumulative I/O write latency | μs | counter |
| 60005 | Cumulative bytes written by I/O operations | Byte | counter |
| 60021 | Cumulative successful MemStore write lock acquisitions | count | counter |
| 60022 | Cumulative failed MemStore write lock acquisitions | count | counter |
| 60023 | Cumulative time spent waiting for MemStore write locks | μs | counter |
| 80057 | Cumulative transaction log bytes | Byte | counter |
Caches, MemStore, and resources
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 120000 | Current location cache memory usage | Byte | gauge |
| 120001 | Current clog cache memory usage | Byte | gauge |
| 120008 | Row cache size | Byte | gauge |
| 120009 | Bloom filter cache size | Byte | gauge |
| 130000 | Active MemStore size | Byte | gauge |
| 130001 | Total MemStore size | Byte | gauge |
| 130002 | Threshold for triggering a major freeze | Byte | gauge |
| 130004 | MemStore size limit | Byte | gauge |
| 140002 | Maximum available memory | Byte | gauge |
| 140003 | Used memory | Byte | gauge |
| 140005 | Maximum CPU cores available to the tenant × 100 | 0.01 core | gauge |
| 140006 | CPU cores actually used by the tenant × 100 | 0.01 core | gauge |
| 140013 | Cumulative CPU time of tenant threads as recorded by cgroup | μs | counter |
Database time model
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 200002 | Cumulative CPU time for foreground database activity (DB CPU) | μs | counter |
| 200006 | Cumulative CPU time of background processes | μs | counter |
| 200010 | Cumulative foreground non-idle wait time | μs | counter |
| 200013 | Cumulative background non-idle wait time | μs | counter |
Shared Storage microblock cache I/O
The following metrics already represent IOPS or throughput, so there is no need to apply rate().
stat_id |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
| 245012 | Current read IOPS for Shared Storage microblock cache common_io |
count/s | gauge |
| 245013 | Current read throughput for Shared Storage microblock cache common_io |
Byte/s | gauge |
| 245014 | Current write IOPS for Shared Storage microblock cache common_io |
count/s | gauge |
| 245015 | Current write throughput for Shared Storage microblock cache common_io |
Byte/s | gauge |
SQL Audit monitoring metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
sql_audit_slow_sql_total_count |
Cumulative number of records identified as slow SQL by the agent's SQL Audit collection | count | counter |
sql_audit_failed_sql_total_count |
Cumulative number of records with FailCount greater than 0 in the agent's SQL Audit collection |
count | counter |
ODP and Binlog metrics
ODP container and Kubernetes resource metrics
Common labels include obproxy_cluster, namespace, pod, and container, as well as metric-specific device, network interface, and resource labels. The actual labels depend on upstream collection and relabeling by the cloud platform. The semantics below are based on the cAdvisor metric definitions and kube-state-metrics Pod metrics. For the CPU quota unit in microseconds, see the cAdvisor quota discussion.
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
container_cpu_load_average_10s |
Container CPU load average over the last 10 seconds, multiplied by 1000; divide the returned value by 1000 to obtain the actual load average | dimensionless | gauge |
container_cpu_usage_seconds_total |
Cumulative CPU time consumed by the container | s | counter |
container_fs_reads_bytes_total |
Cumulative bytes read from the container's file system | Byte | counter |
container_fs_reads_total |
Cumulative number of completed read operations on the container's file system | count | counter |
container_fs_writes_bytes_total |
Cumulative bytes written to the container's file system | Byte | counter |
container_fs_writes_total |
Cumulative number of completed write operations on the container's file system | count | counter |
container_memory_cache |
Container page cache memory | Byte | gauge |
container_memory_usage_bytes |
Current container memory usage | Byte | gauge |
container_network_receive_bytes_total |
Cumulative bytes received over the container's network | Byte | counter |
container_network_transmit_bytes_total |
Cumulative bytes transmitted over the container's network | Byte | counter |
container_spec_cpu_quota |
CPU time quota allowed per CFS period | μs/period | gauge |
container_spec_memory_limit_bytes |
Configured memory limit for the container | Byte | gauge |
kube_pod_container_resource_limits |
Configured resource limits for containers in a Pod | Determined by unit: core for CPU, Byte for memory | gauge |
ODP performance metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
odp_transaction_total |
Cumulative number of transactions processed by OBProxy | count | counter |
odp_sql_request_total |
Cumulative number of SQL requests processed by OBProxy | count | counter |
odp_sql_cost_total |
Cumulative time spent by OBProxy processing SQL requests | ms | counter |
odp_current_session |
Current number of OBProxy connections | count | gauge |
odp_used_connections |
Current number of OBProxy client connections | count | gauge |
odp_new_client_connections |
Cumulative number of client connections that logged in to OBProxy | count | counter |
ODP process metrics
These metrics have the same names as the agent's observer metrics but represent different processes and collection sources. Keep the cluster, node, and process dimensions in queries to avoid combining the two types of processes in a sum.
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
process_cpu_percent |
CPU utilization of the OBProxy process | % | gauge |
process_memory_rss_bytes |
Resident set size (RSS) of the OBProxy process | Byte | gauge |
Binlog metrics
Metric |
Description |
Raw unit |
Value semantics |
|---|---|---|---|
binlog_instance_disk_total_size_mb |
Log disk space used by the Binlog instance | MiB (written as mb in the metric name) | gauge |