[Linux] linux | slow service response, troubleshooting | caused by bandwidth problems | network speed

1. Description

        1. The project uses cloud server, server configuration: 5M bandwidth, 4 cores, 32G, 1T, CentOS7

        2. The CPU, memory, and disk IO have not reached the bottleneck. It is guessed that it is a bandwidth problem.

        3. There are many applications, application + middleware, more than a dozen

        4. The number of people online at the same time is 30+

        5. It has been roughly guessed that it is a bandwidth problem. The document mainly records how to provide evidence to the leader for feedback

2. Evidence collection

1. htop confirms cpu and memory usage

1) Install htop

yum install -y htop

2) Real-time situation

htop

3) Results

Explanation 1: The cpu is not used much

Note 2: The memory is about 76%, not exceeding 85, which is barely enough

~~

2. Bandwidth monitoring

1) Check the network card

ifconfig

2) Use iftop to view directly

      

iftop -n -P -b -i eth0

TX: send

RX: receive

TOTAL: all

cum: total traffic from running iftop to the present time

peak: high peak

rates: average value, the three columns correspond to: the average value of 2s 10s 40s

3) If there is no iftop, install it directly

yum install -y iftop

4) iftop is not obvious, use nload, install nload

yum install -y nload

5) Flow monitoring

nload device eth0

Incoming: incoming traffic, upstream traffic

Outgoing: Outgoing traffic, downlink traffic, cloud server bandwidth limit is downlink

Curr: current bandwidth

Avg: average

as can be seen in other

~~