Trigger an inspection

This topic describes how to trigger an inspection of a cluster by using an API. An inspection can be a basic inspection or a performance inspection.

Call the API

If a password is set, you must pass the authentication. For more information, see API hybrid encryption.

Request path

POST /api/v1/obcluster/inspection

Request parameters

Response

Data structure of DagDetailDTO:

Examples

Request example

POST 10.10.10.1:2886/api/v1/obcluster/inspection

{
    "scenario": "basic"
}

Response example

{
    "successful": true,
    "timestamp": "2025-10-14T15:14:01.277047628+08:00",
    "duration": 50,
    "status": 200,
    "traceId": "e6f64ba7f54e7a47",
    "data": {
        "id": "11001",
        "dag_id": 1001,
        "name": "Cluster inspection",
        "stage": 1,
        "max_stage": 3,
        "maintenance_type": 0,
        "maintenance_key": "",
        "state": "READY",
        "operator": "RUN",
        "start_time": "0001-01-01T00:00:00Z",
        "end_time": "0001-01-01T00:00:00Z",
        "additional_data": null,
        "nodes": null
    }
}

Overview

Basic inspection tasks

Basic inspection tasks focus on the stability and availability of the OceanBase cluster. They include error checks, log checks, disk checks, table checks, tenant checks, cluster checks, archive checks, network checks, and system checks. These inspection items aim to promptly identify potential issues that may lead to cluster failures or data loss, ensuring the cluster runs stably. Basic inspections cover health checks from the system level, such as kernel parameters and file system configurations, to the application level, such as deadlock detection, leader election, and memory usage.

List of basic inspection tasks

• bugs.*

• err_code.*

• clog.*

• ls.*

• log.log_size

• log.log_size_with_ocp

• disk.data_disk_full

• disk.disk_full

• disk.disk_hole

• disk.clog_abnormal_file

• disk.sstable_abnormal_file

• disk.mount_disk_full

• disk.xfs_repair

• table.information_schema_tables_two_data

• table.auto_split_error

• tenant.tenant_min_resource

• tenant.writing_throttling_trigger_percentage

• tenant.ddl_operation_table_size

• tenant.parameters_default

• tenant.tenant_locality_consistency_check

• tenant.max_stale_time_for_weak_consistency

• cluster.auto_increment_cache_size

• cluster.datafile_next

• cluster.data_path_settings

• cluster.deadlocks

• cluster.freeze_trigger_percentage

• cluster.global_indexes_too_much

• cluster.logons_check

• cluster.ls_number

• cluster.major

• cluster.major_suspended

• cluster.memory_chunk_cache_size

• cluster.memory_limit_percentage

• cluster.memory_limit_vs_phy_mem

• cluster.memstore_limit_percentage

• cluster.mod_too_large

• cluster.no_leader

• cluster.ob_enable_plan_cache_bad_version

• cluster.ob_query_timeout

• cluster.observer_not_active

• cluster.observer_port

• cluster.optimizer_better_inlist_costing_parmmeter

• cluster.part_trans_action_max

• cluster.resource_limit_max_session_num

• cluster.server_permanent_offline_time

• cluster.session_limit

• cluster.sys_log_level

• cluster.sys_obcon_health

• cluster.table_history_too_many

• cluster.task_opt_stat_gather_fail

• cluster.tenant_locks

• cluster.tenant_memory_tablet_count

• cluster.tenant_number

• cluster.upgrade_finished

• cluster.upper_trans_version

• cluster.zone_not_active

• cluster.core_file_find

• network.local_ip_check

• network.TCP-retransmission

• system.aio

• system.arm_smmu

• system.check_command

• system.check_system_language

• system.clock_source

• system.clock_source_check

• system.core_pattern

• system.dependent_software

• system.dependent_software_swapon

• system.getenforce

• system.instruction_set_avx

• system.kernel_bad_version

• system.mount_options

• system.parameter

• system.parameter_ip_local_port_range

• system.parameter_tcp_rmem

• system.parameter_tcp_wmem

• system.ulimit_parameter

• version.*

Performance inspection tasks

Performance inspection tasks focus on optimizing the performance and resource utilization efficiency of the OceanBase cluster. They include index optimization, query performance, compression algorithm configuration, network performance, CPU performance, and disk I/O performance. These inspection items aim to identify configuration issues or resource bottlenecks that may affect system performance, helping to optimize the cluster's running efficiency. Performance inspections cover various dimensions related to performance, such as query execution parameters, cache configurations, network latency, disk IOPS, and macroblock utilization.

List of performance inspection tasks

• index.global_index_unpartitioned

• cluster.autoinc_cache_refresh_interval

• cluster.clog_sync_time_warn_threshold

• cluster.cpu_quota_concurrency

• cluster.default_compress_func

• cluster.enable_lock_priority

• cluster.large_query_threshold

• cluster.memstore_usage

• cluster.ob_enable_prepared_statement

• cluster.syslog_io_bandwidth_limit

• cluster.task_opt_stat

• cluster.trace_log_slow_query_watermark

• cpu.oversold

• disk.disk_iops

• tenant.tenant_threshold

• network.log_easy_slow

• network.network_speed

• network.network_drop

• network.network_speed_diff

• network.network_offset

• network.network_write_cond_wakeup

• system.tcp_tw_reuse

• system.cgroup_version