source ：cnblogs.com/xiaoyuxixi/p/12235979.html
New companies need to be monitored , The interview mentioned Prometheus It is the monitoring solution that the company needs , Of course, I chose to follow the trend .
The main thing I did before was Zabbix, Since the company needs Prometheus, There's no way out. , Only a good comparison , Get to know , After all, the skill is not too much .
But a little deeper , I realized Prometheus The advantages of , Summarize these two monitoring methods .
A brief history of the two monitoring tools
Prometheus
Kubernetes since 2012 Since its open source in, it has become an irresistible leader in scheduling and scheduling in the container field .
Kubernetes yes Google Borg Open source implementation of the system , Corresponding to this Prometheus It is Google BorgMon Open source implementation .
Prometheus By SoundCloud Open source monitoring and alarm system and time series database developed .
Literally ,Prometheus It consists of two parts , One is the monitoring and alarm system , The other is its own time series database （TSDB）.
2016 year , from Google Sponsored Linux The original cloud foundation under the foundation （Cloud Native Computing Foundation） take Prometheus Included in its second largest open source project .
Prometheus It's also very active in the open source community , stay GitHub I have more than 20000 Star, And the system will have a small version update every two weeks , and Prometheus With its “ Elder martial brother ”Kubernetes All have their own aura of cloud origin , Nature can be friendly and cooperative .
Zabbix
Zabbix The official release date can be traced back to 2012 year , Time is better than Prometheus Four years earlier .
Zabbix By Alexei Vladishev Open source distributed monitoring system , It is an enterprise level distributed open source monitoring scheme . Software that can monitor various network parameters and server health and integrity . Use flexible notification mechanisms , Allows users to configure mail based alerts for almost any event .
In this way, we can quickly feed back the problems of the server . Based on stored data , Provides excellent reporting and data visualization capabilities .
Architecture contrast
Prometheus
Prometheus The basic principle of HTTP Periodically grab the status of the monitored component , Any component only needs to provide corresponding HTTP Interface and conform to Prometheus Defined data format , You can access it Prometheus monitor .
Prometheus Server Be responsible for grasping the target regularly Metrics（ indicators ） The data is stored in local storage .
Prometheus One was adopted Pull（ PULL ） How to get data , Not only does it reduce the complexity of the client , The client only needs to collect data , No need to know about the server side , And the server can be more convenient for horizontal expansion .
If the monitoring data reaches the alarm threshold Prometheus Server Will pass HTTP Send the alarm to the alarm module alertmanger, After the alarm is suppressed, the e-mail or webhook.
Prometheus Support PromQL Provide multi-dimensional data model and flexible query , By monitoring the indicators associated with multiple tag The way , Combine and aggregate the monitoring data in any dimension .
Zabbix
Zabbix from 2 Part of the form ,Zabbix Server With optional components Zabbix Agent.Zabbix Server Can pass SNMP,Zabbix Agent,ping, Port monitoring and other methods are provided for remote servers / Monitoring of network status , Data collection and other functions .
It can run in Linux,Solaris,HP-UX,AIX,Free BSD,Open BSD,OS X Platforms such as .
The core components are mainly Agent and Server, among Agent It is mainly responsible for collecting data and sending it to Server/Proxy, besides , To extend the monitor ,Agent It also supports the execution of custom scripts .
Server Mainly responsible for receiving Agent Monitoring information sent , And carry out summary storage , Trigger alarm, etc .
Zabbix Server Store the collected monitoring data in Zabbix Database in .Zabbix Database Support common relational database , If MySQL、PostgreSQL、Oracle etc. , The default is MySQL, And provide Zabbix Web page （PHP To write ） Data query .
Zabbix Because of the use of relational data to store temporal data , Therefore, when monitoring large-scale clusters, it is often difficult to save data .
So from Zabbix 4.2 Support for TimescaleDB Time series database , But the maturity is not high yet .
Comprehensive comparison
The table above , In terms of development language , In order to meet the needs of high concurrency and fast iteration , The development language of the monitoring system has gradually changed from C Language transfer to Go.
Have to say ,Go With simple syntax and elegant concurrency , stay Java Occupy business development ,C In the case of occupying the bottom development , Positioning middleware development requirements accurately , It is widely used in current open source middleware products .
In terms of system maturity ,Zabbix It's an old monitoring system ：Zabbix Is in 1998 It came out in , The system function is relatively stable , Higher maturity .
and Prometheus It was born in recent years , Although the functions are still being updated iteratively , But on the shoulders of giants , In the architecture design, we have used the experience of many old monitoring systems for reference .
In terms of data storage ,Zabbix Using relational database to save , This greatly limits Zabbix Acquisition performance , and Prometheus Self developed a set of high performance time series database , stay V3 Version can reach tens of millions of data storage per second , Expand the storage of historical data by docking with the third-party time series database .
In terms of configuration complexity ,Prometheus There's only one core server Components , One command to start , Compared with , Other system configuration is relatively troublesome .
In terms of community activity , at present Zabbix More active , But it's basically domestic companies that participate in ,Prometheus In this respect, we have an absolute advantage , Although the community is not as active as , But by CNCF Support for , The later development is worth looking forward to .
From the perspective of container support , because Zabbix Appeared earlier , At that time, the container was not born , Naturally, the support for containers is also relatively poor .
and Prometheus Dynamic discovery mechanism , Not only can it support Swarm Native clusters , And support Kubernetes Monitoring of container clusters , It is the best solution for container monitoring at present .
summary
Taken together ,Zabbix More mature , It's faster to get started , But better integration leads to less flexibility , The bigger problem is , As the complexity of monitoring data increases ,Zabbix Further customization is very difficult , Even if customized , I can't use the data collected before （ Problems caused by relational databases ）.
Prometheus Basically the opposite , It's more difficult to get started , But due to the high flexibility of customization , Data also has more aggregation possibilities , It is much less difficult to use after starting Zabbix.
But if you have accumulated technology in the traditional monitoring system , Or we should carefully consider changing the monitoring .
If you're monitoring a physical machine , use Zabbix No problem ,Zabbix In traditional monitoring system , Especially in server related monitoring , Take the absolute advantage of .
Even when the environment does not change very frequently ,Zabbix It's better than Prometheus Well done ; But if it's a cloud environment , Unless it is Zabbix Play very slippery , You can make all kinds of customization , Otherwise, or Prometheus Well , After all, that's what people do .
Prometheus It began to become the standard configuration of leading and container monitoring , And it will be widely used in the foreseeable future .
If it's just going to the monitoring system , Don't hesitate ,Prometheus It must be true .
Recent hot article recommends ：
1.1,000+ Avenue Java Arrangement of interview questions and answers (2022 The latest version )
2. Explode ！Java Xie Cheng is coming ...
3.Spring Boot 2.x course , It's too complete ！
4. Don't write about the explosion on the screen , Try decorator mode , This is the elegant way ！！
5.《Java Development Manual （ Song Mountain version ）》 The latest release , Download it quickly ！
I think it's good , Don't forget to like it + Forward ！
