The APMPlus proprietary monitoring system is an independently developed data collection tool for the Application Performance Management Plus full-trace version. It is responsible for collecting monitoring data from applications and infrastructure, and then reporting the relevant data to the APMPlus server side for visualization and analysis.

Operating system architecture

overall, server-side monitoring adopts a three-layer architecture.

Application probe: The integrated Trace, Log, and Metrics SDKs communicate with the local agent via IPC (Inter Process Communication, that is, Inter-process communication).

Agent: Responsible for the collection of host machine data and container data on the host machine, receiving monitoring data sent by applications, caching the data, and forwarding it to the data center after compression and encryption. Also responsible for periodically synchronizing the dynamic configuration from the data center to the application.

Data center: Responsible for receiving data, cleaning and storing it, ultimately presenting it to users, and providing analysis and alert features.

Java application features

Java applications integrate with the server side using a non-intrusive instrumentation solution. The following introduces the features supported after Java applications are integrated with the server side.

Trace

After the Java application connects to the server side, JavaAgent can automatically collect trace information and generate metric data.

For the currently supported frameworks, refer to Supported Frameworks List. After integration, Metrics will be automatically generated as follows:

Metrics	DescriptionS	Tags
apminsight.service.trace.called.throughput	Number of calls	service: Name of called service resource: Name of the called interface status: return code instance_id: Process ID
apminsight.service.trace.called.latency.us.	Invocation duration	service: Name of the called service resource: Name of the called interface status: return code instance_id: Process ID
apminsight.service.trace.call.throughput	Number of calls to other services	service: Service name resource: interface name status: return code instance_id: Process ID call_service: Service name to call call_resource: Interface name to call db.slow_query: Is it a slow SQL query
apminsight.service.trace.call.latency.us.	Duration of calling other services	service: Name of the called service resource: Name of the called interface status: return code instance_id: Process ID call_service: Service name to call call_resource: Called interface name db.slow_query: Is it a slow SQL query
apminsight.service.performance.cpu.core	Service CPU usage	service: Service name
apminsight.service.performance.cpu_limit.core	Service CPU limits	service: Service name
apminsight.service.performance.rss.mb	Service memory usage	service: Service name
apminsight.service.performance.memory_limit.mb	Service memory limit	service: Service name

Log

After a Java application connects to the server side, it can support invocation chain search and log search features. Currently, integration has been implemented for the following frameworks:

Logback

log4j2

Metrics

After a Java application connects to the server side, the Metrics feature becomes available. For details, refer to apminsight-server-sdk-java.

JVM monitoring

After a Java application connects to the server side, JVM-related monitoring is supported, mainly including the following metrics:

Type	Metrics	DescriptionS	Unit
Heap memory	jvm.init_heap	The initial heap memory of the JVM	Byte
		jvm.used_heap	Used heap memory	Byte
		jvm.max_heap	Maximum heap memory for JVM	Byte
		jvm.committed_heap	JVM committed heap memory	Byte
Non-heap memory	jvm.init_noheap	JVM initial non-heap memory	Byte
		jvm.used_noheap	JVM non-heap memory in use	Byte
		jvm.max_noheap	Maximum JVM non-heap memory	Byte
		jvm.committed_noheap	JVM submitted non-heap memory	Byte
Memory pool - NewGen	vm.init_newgen	JVM memory pool - initial NewGen	Byte
		jvm.max_newgen	JVM memory pool - maximum NewGen	Byte
		jvm.used_newgen	JVM memory pool – used NewGen memory	Byte
		jvm.committed_newgen	JVM memory pool - committed NewGen	Byte
Memory pool - OldGen	jvm.init_oldgen	JVM memory pool - initial OldGen	Byte
		jvm.max_oldgen	Maximum JVM OldGen memory pool	Byte
		jvm.used_oldgen	OldGen used	Byte
		jvm.committed_oldgen	JVM memory pool - committed OldGen	Byte
The memory pool CodeCache	jvm.init_code_cache	JVM memory pool - initial CodeCache	Byte
		jvm.max_code_cache	JVM memory pool: maximum CodeCache	Byte
		jvm.used_code_cache	JVM memory pool - CodeCache used	Byte
		jvm.committed_code_cache	JVM memory pool - submitted CodeCache	Byte
Memory pool - survivor	jvm.init_survivor	JVM memory pool - initial Survivor	Byte
		jvm.max_survivor	JVM memory pool: maximum Survivor space	Byte
		jvm.used_survivor	JVM memory pool - Survivor used	Byte
		jvm.committed_survivor	JVM memory pool - committed Survivor	Byte
Memory pool–MetaSpace	jvm.init_meta_space	JVM memory pool - initial MetaSpace	Byte
		jvm.max_meta_space	JVM Memory pool - Maximum MetaSpace	Byte
		jvm.used_meta_space	JVM memory pool – MetaSpace used	Byte
		vm.committed_meta_space	JVM memory pool - committed MetaSpace	Byte
Memory pool-PermGen	jvm.init_perm_gen	JVM memory pool - initial PermGen size	Byte
		jvm.max_perm_gen	JVM memory pool-Maximum PermGen	Byte
		jvm.used_perm_gen	JVM memory pool - used PermGen	Byte
		jvm.committed_perm_gen	JVM memory pool - committed PermGen	Byte
Class loading	jvm.loaded_class_count	The number of classes currently loaded by the JVM	count
		jvm.total_unloaded_class_count	Total number of unloaded classes in the JVM	count
		vm.total_loaded_class_count	The total number of loaded classes in the JVM	count
Order	jvm.live_count	The number of live JVM threads	count
		jvm.daemon_count	Number of JVM daemon threads	count
		jvm.peak_count	JVM maximum thread count	count
		jvm.runnable_state_thread_count	JVM runnable thread count	count
		jvm.blocked_state_thread_count	Number of blocked threads in the JVM	count
		vm.waiting_state_thread_count	JVM waiting thread count	count
		vm.timed_waiting_state_thread_count	JVM timed waiting thread count	count
GC count	jvm.gc_count_old	JVM GC count - OldGen	count
GC count		jvm.gc_count_new	JVM GC count – new generation	count
GC time	jvm.gc_time_old	JVM GC time - OldGen	ms
GC time		jvm.gc_time_new	JVM GC time - NewGen	ms

Custom instrumentation

After a Java application is integrated with the server side, custom code instrumentation for monitoring is supported. Configure the class names and method names for global and service dimensions on the platform to complete instrumentation of the relevant methods. The instrumentation results will be reported and queried together with the trace.

Supported frameworks

Framework/component	Maven	Version
tomcat	org.apache.tomcat.embed:tomcat-embed-core	7.x~8.x
mysql	mysql:mysql-connector-java	5.x~8.x
oracle	com.oracle:ojdbc14	Versions 10.x and above
redis	redis.clients:jedis	Version 2.x and above
redis		io.lettuce:lettuce-core	Version 5.x and above
Memcached	net.spy:spymemcached	Version 2.x or later
Memcached		com.googlecode.xmemcached:xmemcached	Version 2.x and later
MongoDB	org.mongodb:mongo-java-driver	2.x~3.x
MongoDB		org.mongodb:mongodb-driver-sync	Version 4.x or later
Kafka	org.apache.kafka:kafka-clients	Versions 0.11.0.0 and above
RocketMQ	com.alibaba.rocketmq:rocketmq-client	Versions 3.x and above
RocketMQ		org.apache.rocketmq:rocketmq-client	Version 4.x or later
RabbitMQ	com.rabbitmq:amqp-client	Version 5.x and above
logback	ch.qos.logback:logback-classic	Version 1.x or later
log4j	org.apache.logging.log4j:log4j-core	Version 2.x or later
HttpClient	com.squareup.okhttp:okhttp	version 2.x or later
		com.squareup.okhttp3:okhttp	3.x~4.x
		commons-httpclient:commons-httpclient	Versions 3.x and above
		org.apache.httpcomponents:httpclient	Version 4.x and above
		org.apache.httpcomponents.client5:httpclient5	Version 5.x or later
		org.apache.httpcomponents:httpasyncclient	Version 4.x or later
		org.asynchttpclient:async-http-client	Versions 2.x and above
		io.github.openfeign:feign-core	9.x
		org.springframework.cloud:spring-cloud-starter-feign	1.x~2.x
Rest tempalte	org.springframework:spring-web	Version 4.x or later
Spring	org.springframework:spring-webmvc	3.x~5.x
Spring		org.springframework:spring-context	3.x~5.x

Go application features

The Go application integrates with the server side using an intrusive instrumentation approach. The following introduces the features supported after the Go application is integrated with the server side.

Trace

When integrating Go applications with the server side, a non-intrusive instrumentation solution cannot be used; integration can only be achieved through frameworks. Currently, integration has been implemented for the following frameworks:

gin-gonic/gin

gorm.io/gorm.v1

net/http

For details on how to integrate, refer to the trace feature description. After integration, the following Metrics will be automatically generated:

Metrics	DescriptionS	Tags
apminsight.service.trace.called.throughput	Number of calls	Called service name resource: Name of the called interface status: return code instance_id: Process ID
apminsight.service.trace.called.latency.us.	Invocation duration	service: Name of the called service resource: Name of the called interface status: return code instance_id: Process ID
apminsight.service.trace.call.throughput	Number of calls to other services	service: Service name resource: interface name status: return code instance_id: process IDcall_service: Service name to call Name of the called interface db.slow_query: Is it slow SQL
apminsight.service.trace.call.latency.us.	Time to call other services	service: Name of the service being called resource: Name of the called interface status: return code instance_id: process IDall_service: Service name to call call_resource: Name of the called interface db.slow_query: Is it a slow SQL query
apminsight.service.performance.cpu.core	Services CPU usage	service: Service name
apminsight.service.performance.cpu_limit.core	Service CPU limits	service: Service name
apminsight.service.performance.rss.mb	Service memory usage	service: Service name
apminsight.service.performance.memory_limit.mb	Service memory limits	service：Service name

Log

After Go applications are integrated with the server side and adapted to the Trace and Log frameworks, features such as call chain search and log search can be implemented. Currently, integration has been performed for the following frameworks:

sirupsen/logrus

For specific integration methods, refer to Log feature description. The following Metrics will be automatically generated after integration:

Metrics	DescriptionS	Tags
apminsight.service.log.throughput	Service log entries	service: Service name log_level: Log level

Metrics

Go SDK provides the following interfaces

Refer to metrics feature description for specific integration methods.

Supported frameworks list

Framework	Version
gin-gonic/gin	1.x
gorm.io/gorm.v1	1.x
net/http	-
sirupsen/logrus	1.x