All Products
Search
Document Center

Application Real-Time Monitoring Service:Application Monitoring metrics

Last Updated:Nov 26, 2025

This topic describes the common metrics in ARMS Application Monitoring. You can use these metrics to create custom Grafana dashboards.

Important

Applications connected through Managed Service for OpenTelemetry support only business metrics. Other metrics, such as JVM metrics and system metrics, are not supported.

Business metrics

Common dimensions

Dimension name

Dimension key

Service name

service

Service PID

pid

Machine IP

serverIp

Interface

rpc

Metric list

All access types include the following metrics. When you run a query, replace $callType with a specific access type. For more information about access types, see Service access types and available dimensions.

For example, to query the number of requests for an HTTP service, change arms_$callType_requests_count to arms_http_requests_count.

Metric name

Metric

Metric type

Collection interval

Unit

Dimensions

Request count

arms_$callType_requests_count

Gauge

15 s

None

The dimensions vary based on the service access type. For more information, see Service access types and available dimensions.

Error request count

arms_$callType_requests_error_count

Gauge

15 s

None

Request duration

arms_$callType_requests_seconds

Gauge

15 s

Seconds

Slow request count

arms_$callType_requests_slow_count

Gauge

15 s

None

Request duration quantiles

arms_$callType_requests_latency_seconds

Summary

15 s

Seconds

This dimension is available only when the service access type is HTTP and quantile statistics are enabled. To enable quantile statistics, see Advanced Settings.

Quantile dimension:

  • 0.5

  • 0.75

  • 0.90

  • 0.99

Note

Except for quantile metrics, the preceding metrics are all of the Gauge type. This means the value of each data point represents the cumulative total value within the collection interval. This is different from metrics generated by open source frameworks. For example, to calculate the average queries per second (QPS) over one minute, the Prometheus Query Language (PromQL) expression for an ARMS metric is sum_over_time(arms_$callType_requests_count[1m])/60. For an open source framework, the expression is typically rate(http_server_requests_count[1m]).

Metric storage instance description

Business metrics are stored in the detailed metric library (metricstore-apm-metrics-detail).

Aggregated business metrics

  • Business metrics create a separate metric for each call type. This makes the PromQL expression long and complex when a single application has multiple call types, such as HTTP and Dubbo.

  • Business metrics record all observation dimensions. However, not all dimensions are needed in some statistical scenarios. This causes poor performance when you directly query business metrics in some scenarios.

To resolve these issues, ARMS provides aggregated business metrics.

Metric description

Aggregated business metrics are classified into the following categories:

  • General

    Records the request count, error count, slow request count, and average request duration for all access types.

  • Database

    Records the request count, error count, slow request count, and average request duration for database access types.

  • SQL class

    Records the request count, error count, slow request count, and average request duration for database access types. Compared with the database category, this category adds the SQL dimension.

  • Exception

    Records the request count and average request duration for all access types when an exception occurs.

  • Status code

    Records the request count for different status codes in HTTP scenarios.

  • Quantile

    Records the request duration quantiles for all access types.

Except for the quantile category, each major category has two types of metrics. One is the normal aggregated metric, named in the xxx_raw format. The other is the reduced-dimension metric, named in the xxx_ign_x_y format, where x and y are the aggregated dimensions excluded from the metric.

Metric type and collection interval

Unless otherwise specified, all aggregated business metrics are of the Gauge type, and the collection interval is 15 seconds.

Common dimensions

The following dimensions exist in every aggregated metric:

Dimension

Description

pid

Application PID

service

Application name

serverIp

Instance IP

source

Metric source:

  • apm

    Indicates an application connected through an ARMS agent.

  • xtrace

    Indicates an application connected through Managed Service for OpenTelemetry.

  • ebpf

    Indicates an application connected through an Application Monitoring eBPF agent.

Metric list

Metric category

Metric name

Metric

Unit

Other dimensions

General

Request count

arms_app_requests_count_raw

Unit

  • callType: Access type

  • rpc: Interface name

  • prpc: Upstream interface name

  • ppid: Upstream application PID

  • parent: Upstream application name

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

arms_app_requests_count_ign_destid_endpoint_parent_ppid_prpc_rpc

Unit

Excludes the destId, endpoint, parent, ppid, prpc, and rpc dimensions.

arms_app_requests_count_ign_destid_endpoint_rpc_prpc

Unit

Excludes the destId, endpoint, rpc, and prpc dimensions.

arms_app_requests_count_ign_destid_endpoint_parent_ppid_prpc

Item

Excludes the destId, endpoint, parent, ppid, and prpc dimensions.

arms_app_requests_count_ign_parent_ppid_prpc_rpc

Unit

Excludes the parent, ppid, prpc, and rpc dimensions.

Error request count

arms_app_requests_error_count_raw

item

  • callType: Access type

  • rpc: Interface name

  • prpc: Upstream interface name

  • ppid: Upstream application PID

  • parent: Upstream application name

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

arms_app_requests_error_count_ign_destid_endpoint_rpc_prpc

Unit

Excludes the destId, endpoint, rpc, and prpc dimensions.

arms_app_requests_error_count_ign_destid_endpoint_parent_ppid_prpc

Unit

Excludes the destId, endpoint, parent, ppid, and prpc dimensions.

arms_app_requests_error_count_ign_parent_ppid_prpc_rpc

Unit

Excludes the parent, ppid, prpc, and rpc dimensions.

arms_app_requests_error_count_ign_parent_ppid_prpc_rpc

Count

Excludes the parent, ppid, prpc, and rpc dimensions.

Slow request count

arms_app_requests_slow_count_raw

unit

  • callType: Access type

  • rpc: Interface name

  • prpc: Upstream interface name

  • ppid: Upstream application PID

  • parent: Upstream application name

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

arms_app_requests_slow_count_ign_destid_endpoint_parent_ppid_prpc_rpc

Item

Excludes the destId, endpoint, parent, ppid, prpc, and rpc dimensions.

arms_app_requests_slow_count_ign_destid_endpoint_rpc_prpc

Unit

Excludes the destId, endpoint, rpc, and prpc dimensions.

arms_app_requests_slow_count_ign_destid_endpoint_parent_ppid_prpc

Item

Excludes the destId, endpoint, parent, ppid, and prpc dimensions.

arms_app_requests_slow_count_ign_parent_ppid_prpc_rpc

Item

Excludes the parent, ppid, prpc, and rpc dimensions.

Request duration

arms_app_requests_seconds_raw

Seconds

  • callType: Access type

  • rpc: Interface name

  • prpc: Upstream interface name

  • ppid: Upstream application PID

  • parent: Upstream application name

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

arms_app_requests_seconds_ign_destid_endpoint_parent_ppid_prpc_rpc

Seconds

Excludes the destId, endpoint, parent, ppid, prpc, and rpc dimensions.

arms_app_requests_seconds_ign_destid_endpoint_rpc_prpc

Seconds

Excludes the destId, endpoint, rpc, and prpc dimensions.

arms_app_requests_seconds_ign_destid_endpoint_parent_ppid_prpc

Seconds

Excludes the destId, endpoint, parent, ppid, and prpc dimensions.

arms_app_requests_seconds_ign_parent_ppid_prpc_rpc

Seconds

Excludes the parent, ppid, prpc, and rpc dimensions.

Database

Database request count

arms_db_requests_count_raw

Unit

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

arms_db_requests_count_ign_rpc

Item

Excludes the interface dimension.

Database request error count

arms_db_requests_error_count_raw

Unit

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

arms_db_requests_error_count_ign_rpc

Unit

Excludes the interface dimension.

Database slow request count

arms_db_requests_slow_count_raw

Item

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

arms_db_requests_slow_count_ign_rpc

Count

Excludes the interface dimension.

Database request duration

arms_db_requests_seconds_raw

Seconds

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

arms_db_requests_seconds_ign_rpc

Seconds

Excludes the interface dimension.

SQL Class

SQL request count

arms_sql_requests_count_raw

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

  • sqlId: SQL statement ID

arms_sql_requests_count_ign_rpc

Excludes the interface dimension.

SQL request error count

arms_sql_requests_error_count_raw

Unit

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

  • sqlId: SQL statement ID

arms_sql_requests_error_count_ign_rpc

Item

Excludes the interface dimension.

SQL slow request count

arms_sql_requests_slow_count_raw

Count

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

  • sqlId: SQL statement ID

arms_sql_requests_slow_count_ign_rpc

Unit

Excludes the interface dimension.

SQL request duration

arms_sql_requests_seconds_raw

Seconds

  • callType: Access type

  • rpc: Interface name

  • endpoint: Database instance address

  • destId: Database name

  • sqlId: SQL statement ID

arms_sql_requests_seconds_ign_rpc

Seconds

Excludes the interface dimension.

Exception

Exception request count

arms_exception_requests_count_raw

Item

  • callType: Access type

  • rpc: Interface name

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

  • excepInfo: Exception encoding method

  • excepType: Exception encoding ID

  • excepName: Exception name

arms_exception_requests_count_ign_destid_endpoint_rpc

Unit

Excludes the destid, endpoint, and rpc dimensions.

arms_exception_requests_count_ign_destid_endpoint

Unit

Excludes the destid and endpoint dimensions.

Exception request duration

arms_exception_requests_seconds_raw

Seconds

  • callType: Access type

  • rpc: Interface name

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

  • excepInfo: Exception encoding method

  • excepType: Exception encoding ID

  • excepName: Exception name

arms_exception_requests_seconds_ign_destid_endpoint_rpc

Seconds

Excludes the destid, endpoint, and rpc dimensions.

arms_exception_requests_seconds_ign_destid_endpoint

Seconds

Excludes the destid and endpoint dimensions.

Status code

Request count by status code

arms_requests_by_status_count_raw

Count

  • rpc: Interface name

  • status status code:

    • 200: HTTP status code is 200

    • 2xx: HTTP status code is greater than 200 and less than 300

    • 3xx: HTTP status code is greater than or equal to 300 and less than 400

    • 4xx: HTTP status code is greater than or equal to 400 and less than 500

    • 5xx: HTTP status code is greater than or equal to 500

arms_requests_by_status_count_ign_rpc

Item

Excludes the interface dimension.

Quantile

Request duration quantiles

Note

Supported by agent V4.x and later.

arms_uni_requests_latency_seconds

  • callType: Access type

  • rpc: Interface name

  • quantile:

    • 0.5: 50th quantile

    • 0.75: 75th quantile

    • 0.90: 90th quantile

    • 0.99: 99th quantile

  • endpoint: When callType is http_client, this indicates the peer interface. In other scenarios, it indicates the remote address.

  • destId: When callType is a database type, this indicates the database name. In other scenarios, it indicates the remote address.

  • excepInfo: Exception encoding method

  • excepType: Exception encoding ID

  • excepName: Exception name

  • Status code:

    • 200: HTTP status code is 200

    • 2xx: HTTP status code is greater than 200 and less than 300

    • 3xx: HTTP status code is greater than or equal to 300 and less than 400

    • 4xx: HTTP status code is greater than or equal to 400 and less than 500

    • 5xx: HTTP status code is greater than or equal to 500

Usage example

How to select a metric to count the requests of all application interfaces using PromQL

  1. Find a metric that provides the request count for interfaces. The general category of metrics meets this requirement.

  2. The statistical result requires only the interface dimension. Other dimensions, such as upstream interface, upstream application, and remote address, are not required. When you select a specific metric, make sure that the metric includes the interface dimension and as few other dimensions as possible.

Based on the preceding analysis, the optimal metric is arms_app_requests_count_ign_destid_endpoint_parent_ppid_prpc.

Metric storage instance description

The following metrics are stored in the core metric library (metricstore-apm-metrics):

  • arms_app_requests_{metric}_ign_destid_endpoint_parent_ppid_prpc_rpc

  • arms_app_requests_{metric}_ign_destid_endpoint_rpc_prpc

  • arms_app_requests_{metric}_ign_destid_endpoint_parent_ppid_prpc

Other aggregated business metrics are stored in the detailed metric library (metricstore-apm-metrics-detail).

JVM metrics

Common dimensions

Dimension name

Dimension key

Service name

service

Service PID

pid

Machine IP

serverIp

Metric list

Metric name

Metric

Metric type

Collection interval

Unit

Dimensions

Total GC count

arms_jvm_gc_total

Counter

15 s

None

Generational GC area:

  • Young: young generation space

  • Old: old generation space

Cause (GC cause) (supported by agent V4.4.0 and later): System.gc(), Heap Dump Initiated GC, Allocation Failure, and more.

Expand to view detailed descriptions of GC causes

  • System.gc(): This cause is triggered when the program explicitly calls System.gc() to request garbage collection.

  • FullGCAlot / ScavengeAlot: These causes may be related to testing or debugging scenarios where frequent full GCs or minor GCs (scavenges) are intentionally triggered.

  • Allocation Profiler: This GC cause may involve memory allocation profiling to better understand memory usage.

  • JvmtiEnv ForceGarbageCollection: Indicates that the Java Virtual Machine Tool Interface (JVMTI) has requested a forced garbage collection.

  • Heap Inspection Initiated GC & Heap Dump Initiated GC: These are typically not regular operations but are used to inspect the heap or obtain a heap dump.

  • WhiteBox Initiated Young/Full GC & Run to Breakpoint: These are triggered by WhiteBox tests, which are typically used for internal JVM testing purposes.

  • No GC: The JVM may trigger a GC based on internal heuristic algorithms, such as predicting heap usage trends in the Concurrent Mark Sweep (CMS) algorithm, rather than a specific condition like out of memory. Therefore, it is displayed as No GC.

  • Allocation Failure: A garbage collection triggered because the JVM could not allocate memory for a new object, indicating potential memory pressure.

  • CodeCache GC Threshold / Aggressive: Involves the code cache, which stores compiled code. A GC may be triggered when the cache reaches a certain threshold or needs to be managed aggressively.

  • Metadata GC Threshold / Clear Soft References: A GC triggered by the metadata space threshold or the clearing of soft references, which can free up memory without affecting application execution.

  • G1 Evacuation Pause / Compaction Pause / Humongous Allocation / Periodic Collection: These are specific GC causes related to the G1 garbage collector, which is optimized for low pause times.

  • Diagnostic Command: Triggered by a diagnostic command, typically for internal monitoring or diagnostic purposes.

  • Shenandoah GC Causes: These are related to the Shenandoah garbage collector, which aims to reduce pause times through concurrent operations.

  • Z Garbage Collector Causes (Timer, Warmup, etc.): These are specific to ZGC (Z Garbage Collector) and focus on performance tuning aspects, such as warmup, allocation rate, and proactive measures.

  • ILLEGAL VALUE: Indicates an illegal or unrecognized garbage collection cause value, acting as a safeguard against unexpected situations.

Total GC duration

arms_jvm_gc_seconds_total

Counter

15 s

Seconds

GC count per interval

arms_jvm_gc_delta

Gauge

15 s

None

GC duration per interval

arms_jvm_gc_seconds_delta

Gauge

15 s

Seconds

JVM thread count

arms_jvm_threads_count

Gauge

15 s

None

State (thread state):

  • Blocked: Blocked state

  • Live: Active state

  • Daemon: Daemon state

  • New: New state

  • Dead-lock: Deadlocked state

  • Runnable: Runnable state

  • Terminated: Terminated state

  • Timed-wait: Timed waiting state

  • Wait: Waiting state

Initial size of the JVM memory region

arms_jvm_mem_init_bytes

Gauge

15 s

Bytes

Area:

  • Heap: Heap area

  • Nonheap: Non-heap area

  • Total: Total

ID (region breakdown):

  • Eden: Eden space

  • Old: Old generation space

  • Survivor: Survivor space

  • Metaspace: The metadata area

  • Code cache: Code cache

  • Compressed class space

  • Total: Total

Maximum size of the JVM memory region

arms_jvm_mem_max_bytes

Gauge

15 s

Bytes

Used size of the JVM memory region

arms_jvm_mem_used_bytes

Gauge

15 s

Bytes

Committed size of the JVM memory region

arms_jvm_mem_committed_bytes

Gauge

15 s

Bytes

Usage ratio of the JVM memory region

arms_jvm_mem_usage_ratio

Gauge

15 s

Ratio (0 to 1)

JVM loaded classes

arms_class_load_loaded

Counter

15 s

None

None

JVM unloaded classes

arms_class_load_un_loaded

Counter

15 s

None

None

JVM buffer pool size

arms_jvm_buffer_pool_total_bytes

Gauge

15 s

Bytes

ID (region):

  • Direct

  • Mapped

Used size of the JVM buffer pool

arms_jvm_buffer_pool_used_bytes

Gauge

15 s

Bytes

Number of JVM buffer pools

arms_jvm_buffer_pool_count

Gauge

15 s

None

Number of open file descriptors

arms_file_desc_open_count

Gauge

15 s

None

None

Ratio of open file descriptors (open/maximum allowed)

arms_file_desc_open_ratio

Gauge

15 s

Ratio (0 to 1)

None

Metric storage instance description

JVM metrics are stored in the core metric library (metricstore-apm-metrics).

System metrics

Common dimensions

Dimension name

Dimension key

Service name

service

Service PID

pid

Machine IP

serverIp

Metric list

Metric name

Metric

Metric type

Collection interval

Unit

CPU idle percentage

arms_system_cpu_idle

Gauge

15 s

Percentage

CPU I/O wait percentage

arms_system_cpu_io_wait

Gauge

15 s

Percentage

CPU system percentage

arms_system_cpu_system

Gauge

15 s

Percentage

CPU user percentage

arms_system_cpu_user

Gauge

15 s

Percentage

System load (1 minute)

arms_system_load

Gauge

15 s

None

Free disk space

arms_system_disk_free_bytes

Gauge

15 s

Bytes

Total disk space

arms_system_disk_total_bytes

Gauge

15 s

Bytes

Disk usage

arms_system_disk_used_ratio

Gauge

15 s

Ratio (0 to 1)

Memory buffer size

arms_system_mem_buffers_bytes

Gauge

15 s

Bytes

Memory cache size

arms_system_mem_cached_bytes

Gauge

15 s

Bytes

Free memory size

arms_system_mem_free_bytes

Gauge

15 s

Bytes

Free swap memory size

arms_system_mem_swap_free_bytes

Gauge

15 s

Bytes

Swap memory size

arms_system_mem_swap_total_bytes

Gauge

15 s

Bytes

Memory size

arms_system_mem_total_bytes

Gauge

15 s

Bytes

Used memory size

arms_system_mem_used_bytes

Gauge

15 s

Bytes

Inbound network traffic size

arms_system_net_in_bytes

Gauge

15 s

Bytes

Outbound network traffic size

arms_system_net_out_bytes

Gauge

15 s

Bytes

Inbound network errors

arms_system_net_in_err

Gauge

15 s

None

Outbound network errors

arms_system_net_out_err

Gauge

15 s

None

Metric storage instance description

System metrics are stored in the core metric library (metricstore-apm-metrics).

Thread pool/Connection pool metrics

Common dimensions

Dimension name

Dimension key

Service name

service

Service PID

pid

Machine IP

serverIp

Thread pool name (supported by agent versions earlier than 4.1.x)

name

Thread pool type (supported by agent versions earlier than 4.1.x)

type

Metric list

Agent versions 4.1.x and later

Thread pool metrics

Metric name

Metric

Metric type

Collection interval

Dimensions

Core thread count

arms_thread_pool_core_pool_size

Gauge

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Maximum thread count

arms_thread_pool_max_pool_size

Gauge

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Active thread count

arms_thread_pool_active_thread_count

Gauge

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Current thread count

arms_thread_pool_current_thread_count

Gauge

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Peak thread count in the thread pool (the highest number of threads since the pool was created)

arms_thread_pool_max_thread_count

Gauge

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Scheduled task count

arms_thread_pool_scheduled_task_count

Counter

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Completed task count

arms_thread_pool_completed_task_count

Counter

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Rejected task count

arms_thread_pool_rejected_task_count

Counter

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Task queue size

arms_thread_pool_queue_size

Gauge

15 s

  • thread_name_pattern: The thread name pattern, such as http-nio-8080-*.

  • thread_pool_usage: The purpose of the thread pool, such as Tomcat, Dubbo, or Undertow.

Connection pool metrics

Metric description

Metric name

Metric type

Collection interval

Dimensions

Connections

arms_connection_pool_connection_count

Gauge

15 s

  • state: Connection state.

    • used

    • idle

  • pool_type: Connection pool type, such as Druid or c3p0.

  • url: Database connection string.

Minimum idle connections

arms_connection_pool_connection_min_idle_count

Gauge

15 s

  • pool_type: Connection pool type, such as Druid or c3p0.

  • url: Database connection string.

Maximum idle connections

arms_connection_pool_connection_max_idle_count

Gauge

15 s

  • pool_type: Connection pool type, such as Druid or c3p0.

  • url: Database connection string.

Maximum connections

arms_connection_pool_connection_max_count

Gauge

15 s

  • pool_type: Connection pool type, such as Druid or c3p0.

  • url: Database connection string.

Pending connection requests

arms_connection_pool_pending_request_count

Counter

15 s

  • pool_type: Connection pool type, such as Druid or c3p0.

  • url: Database connection string.

Agent versions earlier than 4.1.x

Metric name

Metric

Metric type

Collection interval

Dimensions

Thread pool core size

arms_threadpool_core_size

Gauge

15 s

None

Thread pool maximum size

arms_threadpool_max_size

Gauge

15 s

None

Thread pool active size

arms_threadpool_active_size

Gauge

15 s

None

Thread pool queue size

arms_threadpool_queue_size

Gauge

15 s

None

Thread pool current size

arms_threadpool_current_size

Gauge

15 s

None

Number of tasks in different states in the thread pool

arms_threadpool_task_total

Gauge

15 s

Status (task state):

  • Scheduled: Scheduled

  • Completed: Completed

  • Rejected: Rejected

Metric storage instance description

Thread pool and connection pool metrics are stored in the core metric library (metricstore-apm-metrics).

Scheduled task metrics

The following metrics are available only for scheduled tasks.

Common dimensions

Dimension name

Dimension key

Service name

service

Service PID

pid

Machine IP

serverIp

Task ID

rpc

Metric list

Metric name

Metric

Metric type

Collection interval

Unit

Scheduling delay

arms_$callType_delay_milliseconds

Gauge

15 s

ms

Metric storage instance description

Scheduled task metrics are stored in the detailed metric library (metricstore-apm-metrics-detail).

Go runtime metrics

Metric list

Metric name

Metric

Metric type

Collection interval

Application uptime (ms)

arms_golang_runtime_uptime

Int64Counter

15 s

Number of goroutines in the current application

arms_golang_process_runtime_go_goroutines

Gauge

15 s

Heap object memory (bytes)

arms_golang_process_runtime_go_mem_heap_alloc

Gauge

15 s

Unallocated or reclaimed heap memory

arms_golang_process_runtime_go_mem_heap_idle

Gauge

15 s

Used heap memory

arms_golang_process_runtime_go_mem_heap_inuse

Gauge

15 s

Allocated and live heap objects

arms_golang_process_runtime_go_mem_heap_objects

Gauge

15 s

Memory in HeapIdle that has been released to the operating system

arms_golang_process_runtime_go_mem_heap_released

Gauge

15 s

Size of virtual memory requested from the system

arms_golang_process_runtime_go_mem_heap_sys

Gauge

15 s

Number of current live objects

arms_golang_process_runtime_go_mem_live_objects

Gauge

15 s

Number of GC cycles since the program started

arms_golang_process_runtime_go_gc_count

Gauge

15 s

Cumulative stop-the-world garbage collection pause time, which is the unavailable time

arms_golang_process_runtime_go_gc_pause_total_ns

Int64Counter

15 s

GC pause time distribution

arms_golang_process_runtime_go_gc_pause_ns

Int64Histogram

15 s

Metric storage instance description

Go runtime metrics are stored in the core metric library (metricstore-apm-metrics).

Service access types and available dimensions

Client-side types

  • Access types

    • http_client

    • dubbo_client

    • hsf_client

    • dsf_client

    • notify_client

    • grpc_client

    • thrift_client

    • sofa_client

    • mq_client

    • kafka_client

  • Dimensions

    • parent: Upstream service name

    • ppid: Upstream service PID

    • destId: Extended information about the request peer

    • endpoint: Request peer address

    • excepType: Exception ID

    • excepInfo: Exception ID encoding rule

    • excepName: Exception name

    • stackTraceId: Stack trace ID

DB types

  • Access types

    • mysql

    • oracle

    • mariadb

    • postgresql

    • ppas

    • sqlserver

    • mongodb

    • dmdb

  • Dimensions

    • parent: Upstream service name

    • ppid: Upstream service PID

    • destId: Database name

    • endpoint: Database address

    • excepType: Exception ID

    • excepInfo: Exception ID encoding rule

    • excepName: Exception name

    • stackTraceId: Stack trace ID

    • sqlId: SQL statement ID

Server-side types

  • Access types

    • http

    • dubbo

    • hsf

    • dsf

    • user_method

    • mq

    • kafka

    • grpc

    • thrift

    • sofa

  • Dimensions

    • prpc: Upstream interface

    • parent: Upstream service name

    • ppid: Upstream service PID

    • Endpoint: The address of the service.

    • excepType: Exception ID

    • excepInfo: Exception ID encoding rule

    • excepName: Exception name

    • stackTraceId: Stack trace ID

Scheduled task types

  • Access types

    • xxl_job

    • spring_scheduled

    • quartz

    • elasticjob

    • jdk_timer

    • schedulerx

  • Dimensions

    • prpc: Upstream interface

    • parent: Upstream service name

    • ppid: Upstream service PID

    • excepType: Exception ID

    • excepInfo: Exception ID encoding rule

    • excepName: Exception name

    • stackTraceId: Stack trace ID