What are the software operation and maintenance monitoring?

Monitoring is a function that programmers need to rely on when maintaining software operation , Today we will learn about it through case analysis , What are the indicator types of software operation and maintenance monitoring .

1、 indicators

Usually, the monitoring indicators are from System 、 application 、 Business Wait for several dimensions ：

System monitoring ： Mainly to monitor the physical machine 、 virtual machine 、 How the operating system works , The main indicators include CPU、 Memory 、 disk 、 Network, etc , Other related data include the physical machine run time 、 Operating system version 、 Operating system kernel , These are also some of the basic basis for troubleshooting . Here we need to focus on the network , Microservices are called or called through the network , Once the network goes wrong , The entire microservice cluster is not available , So network monitoring needs to be refined to traffic 、 Data packets 、 Packet loss 、 Misstatement 、 The number of connections, etc .

Application monitoring ： It mainly monitors the operation of the application , Including application run time 、http Service port 、 service url、http Service response code 、http Service response time 、SQL、 A cache hit 、TPS、QPS etc. . about Java application , It also needs to include JVM Operation of the ：JDK edition 、 Memory usage ( Heap memory 、 Non heap memory, etc )、GC etc. Java Virtual machine operation .

Business monitoring ： It mainly monitors the implementation of some core businesses , It's intrusive to the business , The indicators of each service are different , Different monitoring methods are different , It's usually buried code . For example, monitoring login registration 、 Commodity information 、 Inventory 、 Place an order 、 payment 、 Delivery and other business .

2、 health

The general health check-up is done by a heartbeat test , There are usually two kinds of ：

One is to establish TCP link , perform ping/pong call . This way needs to be established with the monitoring system in the service TCP link , Monitoring components need to be embedded in the service , There's an intrusion into the service . But because of its high efficiency , And it's targeted , There will be no underreporting .

One is to listen to the service port , In this way, you only need to add monitoring plug-ins in the container or virtual machine , There's no intrusion into the service , But since port availability and service availability are not a concept , So there will be cases of underreporting .

3、 Exception collection

There are two kinds of anomalies , Logical and behavioral anomalies . Logical exception means that there is abnormal logic in the code , For example, the common NPE; Abnormal behavior refers to the unexpected user behavior , Both of these situations are harmful to the system . So we need to collect these anomalies , And be able to locate the location of the exception . The purpose of abnormal information collection is to locate the problem , So the information reported must be comprehensive and easy to locate . Therefore, it is necessary to protect the exception code in the reported information , You can customize a certain length of string , Easy to position . And then we have to report the parameters , Used to restore the site . And report abnormal information , Used to analyze anomalies .