Shulou(Shulou.com)06/03 Report--

Preface

When we are doing activities similar to Singles Day holiday, the boss will come and ask us these questions.

1. How many visits can the online service withstand 2. How many visits can a single server withstand? 3. Do you need an extra machine? How many machines need to be added?

At this time, it reflects the importance of capacity planning. So what exactly is capacity planning?

Capacity planning is based on current performance as a baseline to determine what and when you need it

Capacity VS performance: determine what a car can hold capacity: determine how many cars are needed

Capacity planning can be broken down into the following four steps

Clear objectives, collection indicators, trends, forecast capacity deployment, clear goals

You should not start capacity planning until you have a clear understanding of site requirements.

Generally speaking, each service has a committed quality of service, so we need to make capacity planning and hardware investment according to this goal.

For example, website A can withstand 3000 QPS and the response time is less than 200 ms.

SLA (Service level Agreement)

Collection index

Capacity planning is not recommended if you do not know the range of capacity that the current service can bear.

We need to test to understand the data metrics of the current service. If you don't test your service limit, it won't work out.

The test is mainly divided into the following two steps

Main functions of the test server: business dimension QPS, TPS test server hardware resources: CPU, memory, hard disk, network trend prediction

Forecasting capacity is a continuous process, which needs to be accurately predicted by mathematics and intuition.

The whole process is as follows

First of all, determine the resource metrics, such as memory consumption, hard disk consumption, and CPU consumption, which actually restrict access to the resources you have: for example, total memory size, total hard disk size, etc. Finally, the specific capacity exceeds the time point according to the monitoring trend chart (Zabbix, Metrics): for example, when the memory will be exhausted and deployed.

Once you have determined how much capacity is needed in the future to meet your business needs, you can start new equipment and deploy it.

Minimize deployment time through automated deployment tools (Ansible, Salt)

Art https://book.douban.com/subject/4200645/ with reference to Web capacity Planning

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.