First, let's figure out a few things:
1. What is the Monitored object?
2. What are the monitoring metrics?
3. What are the dimensions for monitoring?
Monitored object
1. User-side monitoring: monitoring of service functions provided to users
2. interface monitoring: RPC interface monitoring on which business functions depend
3. Resource Monitoring: monitoring of resources on which interfaces depend
4. Basic monitoring: monitoring of the health status of the server itself
Monitoring metrics:
1. Request volume: 1. Real-time request volume (QPS); 2. Count Request volume (PV)
2. Response Time: It is reflected by the average time consumption within a period of time. It can be 0 ~ 10 ms, 10 ms ~ 50 ms, 50 ms ~ 100 ms, Ms ~ The five intervals above 500 ms and Ms are defined based on these five intervals.
3. Error Rate: Percentage of call errors to the total number of calls
Monitoring dimension
1. Global Dimension: an overall understanding of all monitoring objects
2. Data Center dimension: the same monitoring object may be different for different data centers
2. Single-host dimension: the same monitoring object may be different for different hosts
3. time dimension: the same monitoring object may be different for the same time every day (impact of various business activities). It can be defined one day ago, one week ago, one month ago, wait three months ago.
4. Core Dimensions: separate core business and non-core business for monitoring
Summary of microservice calls