Shulou(Shulou.com)06/02 Report--

This article introduces the relevant knowledge of "what are the performance indicators that Java programmers must know?" in the operation of actual cases, many people will encounter such a dilemma, so let the editor lead you to learn how to deal with these situations. I hope you can read it carefully and be able to achieve something!

1. Response time and throughput

According to the response time of the application, you can know the time it takes for the program to complete the transfer of data. It can also be viewed at the HTTP request level or at the database level. For those slow queries, you need to make some optimizations to shorten the time. Throughput is another measure of data transmission, which refers to the number of customer requests processed by the system per unit time.

We can use APMs (such as New Relic or AppDynamics) to measure these metrics. Using these tools, you can directly compare the average response time in the main report dashboard with yesterday's or even last week's. This helps us to see if the new deployment will affect our application. You can see the percentage of network traffic and measure how long it takes HTTP to complete the request.

Recommended tools:

AppDynamics

New Relic

Ruxit

New Relic report: Web Transport percentage and Throughput

two。 Average load

The second widely used indicator is the average load. It is customary to divide the average load into these three steps, namely, the 5th minute, the 15th minute and the first minute. Make sure the quantity is lower than the number of cores of the machine. Once the number of cores is exceeded, the machine will run under pressure.

In addition to simply measuring CPU usage, you need to pay attention to how many processes are in each kernel's queue. With 100% kernel utilization, only 1 task in the queue is very different from 6 tasks. Therefore, the average load cannot only consider CPU utilization.

Recommended tools:

Htop

3. Error rate

Most developers judge the error rate based on the percentage of total HTTP transmission failures. But they overlooked something deeper: the error rate of specific transmissions. This directly affects the health of your application. This shows the errors in the code method and the number of errors or exceptions that have occurred.

But the simple error rate data does not help us much. The most important thing is that we should find their root causes and solve the problems. As Takipi runs, we need to look for clues in the log files. You can find all the information about the status of the server, including stack traces, source code, and variable values.

Recommended tools:

Takipi

4.GC rate and pause time

Abnormal behavior garbage collector application throughput and response time to take deep diving one of the main reasons. The key to understanding the frequency and duration of GC pauses is to analyze GC log files. To analyze them, you need to collect GC logs and JVM parameters. You should pay attention to how the data of different indicators affect each other.

Recommended tools:

JClarity Censum

GCViewer

5. Business indicators

The performance of the application does not depend entirely on response time and error rate. Business indicators are also on the one hand, such as revenue, number of users.

Recommended tools:

Grafana

The ELK stack

Datadog

Librato

6. Uptime and service running status

This metric lays the foundation for the performance of the entire application. Not only can it be used as a reminder, but it also allows you to define the SKA over a period of time. We can use the servlet function of Pingdom to do a health check. We can find all the transmissions of the application, including the database and S3.

Recommended tools:

Pingdom

7. Log size

The log has a drawback, it is always increasing. When your server starts full of garbage, everything slows down. Therefore, we need to pay close attention to the log size.

At present, the common solution is to use logstash to divide usage logs and send and store them in Splunk, ELK, or other log management tools.

Recommended tools:

Splunk

Sumo Logic

Loggly

This is the end of the content of "what are the performance metrics that Java programmers must know?" Thank you for reading. If you want to know more about the industry, you can follow the website, the editor will output more high-quality practical articles for you!

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.