Service metrics

更新时间:
复制 MD 格式

The application monitoring feature of Real-time Monitoring Service (RMS) allows you to analyze applications. You can view and analyze all application monitoring data on the Application Monitoring page. You can aggregate application-related service metrics, such as Error, Service, SAL, CAL, and DAL. Then, you can perform comparative analysis on the metric data based on spatial distribution and temporal distribution of Internet data centers (IDCs), logical data centers (LDCs), and single machines.

In the left-side navigation pane of the RMS console, click Application Monitoring. On the Application Monitoring page, click the My Applications or All Applications tab. Click an application to go to the monitoring details page of the application.

When application monitoring results are displayed from the service dimension, you can configure and monitor the following metrics from the IDC, LDC, or single machine perspectives.

Metric

Description

Configuration Requirement

cal

The metrics related to cache access. The related metrics are used to monitor the access of the current application to the cache.

By default, the metric data is collected from the /home/admin/logs/tracelog/zdal-tair-stat.log file in the Scalable Open Financial Architecture (SOFA) framework. If you use the custom configuration, the metric data can be collected from custom log files. The custom log files must record the following information: data source, results, time consumed, and count.

Sample log: 2015-04-19 21:56:16.627,Test,Tair,1,0,Y,T 2015-04-19 21:56:16.628,Test,Tair,1,20,N,F

The sample log records the following information:

• Time when the log is generated

• Name of the current application

• Logical Tair name

• Number of requests within the current period of time

• Total time consumed by requests within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

error

The number of application errors. The metric is used to monitor the number of error logs in the current application and show the trend of errors. On the Application Monitoring page, you can click an error to view related logs.

By default, the metric data is collected from the /home/admin/logs/stderr.log or /home/admin/logs/{app}/common-error.log file in the SOFA framework. If you use the custom configuration, the metric data can be collected from custom log files. The number of rows in the log files is counted by minute.

pv

The number of page views. The metric is used to monitor the number of calls to pages in the current application.

By default, the metric data is collected from the /home/admin/logs/tracelog/sofa-mvc-stat.log file in the SOFA framework. If you use the custom configuration, the metric data can be collected from custom log files. The custom log files must record the following information: URL, results, time consumed, and access count.

Sample log: 2014-09-01 00:03:22.559,tbapi,http://tbapi.alipay.com/trade/batch_payment.htm,GET,2,11,Y,F

The sample log records the following information:

• Time when the log is generated

• Name of the current application

• Request URL

• Request method

• Number of requests within the current period of time

• Total time consumed by requests within the current period of time

• Request result (A result code that starts with 1 or 2 indicates success, a 302 redirect indicates success, and other result codes indicate failure.)

• Whether stress testing is performed

service

The metric is used to monitor the access of other applications to the current application. In a service-oriented architecture (SOA), complex service dependencies exist between systems. The Service metric indicates the statistical value of the number of calls from other systems to the services exposed by a specific system.

By default, the metric data is collected from the /home/admin/logs/tracelog/rpc-server-stat.log file in the SOFA framework. If you use the custom configuration, the metric data can be collected from custom log files. The custom log files must record the following information: source application, interface, method, results, time consumed, and count.

Sample log: 2014-06-19 17:14:02.186,client,client,com.alipay.cloudenginetest.services.SofaApiWebReferenceLocalFalseTrService:1.0,service_method,1,7,Y,T,GZ00B 2014-06-19 17:15:02.190,client,client,com.alipay.cloudenginetest.services.SofaApiWebReferenceLocalFalseTrService:1.0,service_method,1,12,Y,F,GZ00B

The sample log records the following information:

• Time when the log is generated

• Name of the source application that calls the current application

• Name of the current application

• Service name

• Method name

• Number of times the current application is called within the current period of time

• Total time consumed by requests within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

• Zone where calls are sent

sal

The metrics related to the Service Access Layer (SAL) are used to monitor the calls from the current application to other services. In an SOA, complex service dependencies exist between systems. The Service metric indicates the statistical value of the number of calls from a system to the services exposed by other systems.

By default, the metric data is collected from the /home/admin/logs/tracelog/rpc-client-stat.log file in the SOFA framework. If you use the custom configuration, the metric data can be collected from custom log files. The custom log files must record the following information: IDC, destination application, interface, method, results, time consumed, and count.

Sample log: 2014-06-19 17:14:02.186,client,client,com.alipay.cloudenginetest.services.SofaApiWebReferenceLocalFalseTrService:1.0,service_method,1,79,Y,T,RZ00B 2014-06-19 17:09:02.146,client,,com.alipay.core.service.lazy.LazyPublisherWsService:1.0,service_method,1,26,N,F,RZ00B

The sample log records the following information:

• Name of the current application

• Name of the destination application called by the current application

• Service name

• Method name

• Number of calls within the current period of time

• Total time consumed by calls within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

• Zone where calls are received

ceThread

The number of threads that are running when container threads are busy.

Custom log files of the applications are configured on CE Thread container threads are busy. Sample log: 2014-07-10 00:01:01,316 [CE-RES-MONITOR] - (0,1,79,80,) 2014-07-10 00:01:31,316 [CE-RES-MONITOR] - (0,0,80,80,)

Format: blockQueue active idle poolSize

dal

The metrics related to the Data Access Layer (DAL) are used to monitor the access of applications to databases. Most application systems access databases. The DAL metric indicates the statistical value of the number of calls from a system to databases.

By default, the metric data is collected from the /home/admin/logs/tracelog/zdal-db-stat.log file in the SOFA framework. A statistical log is generated once a minute. If you use the custom configuration, the metric data can be collected from custom log files. The custom log files must record the following information: data source, database, table, results, time consumed, and count.

Sample log: 2015-05-11 20:03:30.487,Test,DataSourceName,LogDatabaseName,tableName,mysql,2,27,N,F 2015-05-11 20:03:30.487,Test,DataSourceName,LogDatabaseName,tableName,mysql,1,0,Y

The sample log records the following information:

• Time when the log is generated

• Name of the current application

• Name of the data source

• Name of the logical database

• Table name

• Database type

• Number of requests within the current period of time

• Total time consumed by requests within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

sql

The metric is used to monitor the execution of SQL statements when the current application accesses databases, including identifying slow SQL queries.

By default, the SOFA framework uses the /home/admin/logs/tracelog/zdal-db-digest.log file. If you use the custom configuration, the metric data can be collected from custom log files. The custom log files must record the following information: database, SQL, time consumed, and transaction ID.

flowlimit

Throttling

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/{app}/guardian/guardian-limit-stat.log.

msgpub

Message publishing

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/tracelog/msg-pub-stat.log.

Sample log: 2018-12-07 16:21:40.168,,TP-openmq-chengyi-1,MQ-PRODUCER-GROUP,8,307,Y,F

The sample log records the following information:

• Time when the log is generated

• Name of the current application

• Topic name

• ID of the publishing group

• Number of messages published within the current period of time

• Messages sending within the current period of time

• Total time consumed

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

msgsub

Message subscription

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/tracelog/msg-sub-stat.log.

Sample log: 2018-12-07 16:23:54.187,,TP-openmq-chengyi-1,S_openmq_chengyi_0,16,67,N,FF

The sample log records the following information:

• Time when the log is generated

• Name of the current application

• Topic name

• ID of the subscription group

• Number of messages received within the current period of time

• Message reception within the current period of time

• Total time consumed

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

servicePressure

Stress testing on calls to an application service by external applications

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/rpc-server-stat.log.

salPressure

Stress testing on calls to external services

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/rpc-client-stat.log.

pvPressure

Stress testing on page views

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/sofa-mvc-stat.log.

dalPressure

Stress testing on calls to database services

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/zdal-db-stat.log.

calPressure

Calls to the cache service

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/zdal-tair-stat.log.

msgpubPressure

Stress testing on messages publishing

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/msg-pub-stat.log.

msgsubPressure

Stress testing on messages subscription

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/tracelog/shadow/msg-sub-stat.log.

mosnservice

Calls to MOSN application services

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/tracelog/mosn/rpc-server-stat.log.

Plug-in type: SPM

A log records the following information:

• Time when the log is generated

• Name of the source application that calls the current application

• Name of the current application

• Service name

• Method name

• Number of times the current application is called within the current period of time

• Total time consumed by requests within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

• Zone where calls are sent

mosnsal

Calls to external application services by MOSN

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/tracelog/mosn/rpc-client-stat.log.

Plug-in type: SPM

A log records the following information:

• Time when the log is generated

• Name of the current application

• Name of the destination application called by the current application

• Service name

• Method name

• Number of calls within the current period of time

• Total time consumed by calls within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

• Zone where calls are received

mosnmsgsub

Subscription to MOSN messages

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/tracelog/mosn/msg-sub-stat.log.

mosnmsgpub

Publishing of MOSN messages

The metric data is collected from the log file configured in the framework. The log file is /home/admin/logs/tracelog/mosn/msg-pub-stat.log.

Plug-in type: SPM

A log records the following information:

• Time when the log is generated

• Name of the current application

• Topic name

• Event code

• ID of the publishing group

• Number of messages published within the current period of time

• Total time consumed for publishing messages within the current period of time

• Result (Y/N)

• Whether full-link stress testing is performed (T/F)

mosnerror

Number of MOSN errors

The metric data is collected from the log files configured in the framework or custom log files. The log files configured in the framework are /home/admin/logs/mosn/default.log, /home/admin/logs/mosn/antvip.error.log, /home/admin/logs/mosn/drm.error.log, /home/admin/logs/mosn/endpoint.error.log, /home/admin/logs/mosn/registry.error.log, /home/admin/logs/mosn/zoneclient.error.log, and /home/admin/logs/mosn/alert.mirror.log.

Plug-in type: ERROR

mosnErrorCode

Number of MOSN error codes

The metric data is collected from the log file configured in the framework or custom log files. The log file configured in the framework is /home/admin/logs/mosn/alert.default.log.

Plug-in type: MM

odpstat

The metric is used to monitor the OceanBase proxy.

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/obproxy/log/obproxy_stat.log.

Plug-in type: MM

A log records the following information:

• Time when the log is generated

• Name of the current application

• Name of the data source (trade_ds) (Reserved. In V1.x, the name is fixed. In V2.x, the name of the logical database is used.)

• Name of the physical database (master00), which is replaced with the schema information (cluster:tenant:database)

• Database type (OceanBase or RDS)

• SQL type (CRUD)

• Execution result (success or failed)

• Error code (which is left empty when the execution result is success)

• Total number of requests

• Number of requests that consume 30 ms to 100 ms

• Number of requests that consume 100 ms to 500 ms

• Number of requests that consume more than 500 ms

• Total execution time that includes the time consumed for executing internal SQL statements (Unit: ms)

• Pre-execution time

• Database execution time

odperror

The number of Open Database Proxy (ODP) exceptions.

The metric data is collected from the log file configured in the framework. The metric data cannot be collected from custom log files. The log file is /home/admin/logs/obproxy/log/obproxy_error.log.

Plug-in type: ERROR