Shulou(Shulou.com)05/31 Report--

This article introduces the relevant knowledge of "what are the most commonly used performance indicators in server monitoring system". In the operation process of actual cases, many people will encounter such difficulties. Next, let Xiaobian lead you to learn how to deal with these situations! I hope you can read carefully and learn something!

PV

PV stands for Page View. Each time a user visits a page through a browser, the application server generates a request,

It is recorded as a PV. Taobao performance test environment, this concept will be extended, the system actually handles a request, depending on

for a PV. That is, the concept of PV also applies to interfaces.

PV statistics can generally be obtained by monitoring buried points or statistical access logs.

There is a special case of PV called PeakPV, which refers to the peak PV value reached during the day.

Through some monitoring systems, statistical data can also be visually seen.

QPS/TPS

QPS/TPS is the number of requests/transactions that the system can process per second, or throughput. In web applications we are more concerned with the number of requests that web applications can process per second. This is an important indicator of system performance.

QPS (TPS)= Number of complications/average response time.

QPS statistics can be obtained by dividing the PV amount of the access log statistics corresponding to the corresponding time by the corresponding time. It can be obtained through tool testing in performance testing.

QPS corresponding to PV during peak period is usually counted frequently.

ResponseTime ResponseTime

Response time (RT) refers to the time elapsed from the time when the client sends a request to the time when the client receives the response result returned from the server. The response time consists of three parts: request sending time, network transmission time and server processing time.

LOAD

Average system load, defined as the average number of processes in the running queue during a given time interval. If a process is full

It is in the run queue if:

- It is not waiting for the result of an I/O operation

- it does not actively enter the wait state (i.e. it does not call 'wait')

- Not stopped (e.g. waiting for termination) 1

The ideal index value for this load value is CPU number * core number *0.7. If it exceeds this value for a long time, it is necessary to be vigilant against the system.

CPU resources

CPU resources here refers to the CPU resource utilization rate of the application service system. CPU resource is an important parameter to judge whether the system processing capacity and application running stability.

JVM GC and FullGC

For Java application performance indicators must be indispensable GC related indicators. In general, our applications should avoid FGC as much as possible. Because FGC does a complete garbage cleanup, which can make applications run slowly, you need to avoid FGC by setting appropriate JVM parameters and GC policies. Common metrics monitored are GC counts and response time.

Commonly used performance indicators are memory occupation, disk io and other indicators, which are not listed here.

The concepts and statistical methods of some performance indicators have been introduced above, and some important connections and differences between several of them are discussed below.

1. capacity prediction

For the system we design, we definitely need to test how much traffic we can receive from users before going live. That is, we want to assess whether our system can sustain the maximum daily PV when it comes. But how to evaluate it? Do you want to create a maximum daily pv scenario to test it? In fact, according to the existing experience and data, we can summarize the relationship between peak QPS and daily pv.

We can see from the daily QPS and PV statistical charts that the daily curves are basically consistent. Through mathematical modeling, we can find that peak QPS per server =( (total PV*80%)/(24*60*60*40%))/number of servers 1. The numbers 80% and 40% are variable parameters, and this formula means that 80% of the total pv is generated in 40% of the time (12 hours). There are different parameters for different scenarios.

In this way, we can obtain its peak QPS through the pressure measurement application, and then calculate the daily PV at the specified peak QPS according to the formula, so as to make capacity prediction.

i.e. PV= QPS *(24*60*60* time percentage)/0.8 * number of machines

2. CPU resource utilization vs. LOAD

According to many people's impression CPU usage and load are statistics of current CPU usage. However, these two indicators are actually quite different.

CPU usage is well understood, that is, the ratio of CPU usage time. CPU load is based on the average length of the queue of tasks waiting to be processed by the CPU over time. This indicator has a higher reference value than cpu occupancy under high load conditions. Because in high load periods, CPU occupancy is basically close to 100%, it does not reflect the degree of machine load. On the contrary, the length of task queue can reflect whether the system load is serious and controllable.

The concept of load can be well understood by using the relationship between road and vehicle in the following figure:

(when the system is a single processor)

When load equals 1, the system is fully loaded, but can meet the current system requirements;

When load is less than 1, the system runs easily.

When load is greater than 1, there are many vehicles waiting to enter the road, just like the task is waiting for CPU processing, at this time CPU occupancy can not distinguish load=1 and load>1 these two cases.

So reading load is helpful for understanding the current operating load of the system.

"What are the most commonly used performance indicators in server monitoring systems" is introduced here. Thank you for reading. If you want to know more about industry-related knowledge, you can pay attention to the website. Xiaobian will output more high-quality practical articles for everyone!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.