> For the complete documentation index, see [llms.txt](https://cp.rohanthe.dev/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://cp.rohanthe.dev/linux-essentials/linux-os-performance.md).

# Linux OS Performance

> Still working on building this page..

Here's a list of commands that I find useful when debugging issues around the performance of Linux VMs when an application is running slow. Through these commands, I try to get a good idea of what might causing the slow performance of the OS and the application. 

**Check the server load** 

```bash
uptime 
#Check load averages to see if anything seems odd.

```

**Check the RAM and CPU usage at a glance with processes**

```bash
htop
#Check if the CPU and the RAM usages are maxing out.
```

**List processes that might seem**

```bash
ps aux | grep <PID/Process Name>
```

**Check the RAM utilization**

```bash
vmstat -Sm 1
#Check r, free, buff, cache, us, sy
```

**Check the disk performance**

```bash
iostat -xmdz 1
#Workload (r/s, w/s, rMB/s, wMB/s), Resulting Performance( avgqu-sz, await, svctm, %util),
```

**Check the processor performance**

```bash
mpstat -P ALL 1
#Look for hot CPUs by checking the %idle time.
```

**Check the amount of free and cached memory**

```bash
free -m
#Check block device I/O cache (Buffers) and virtual page cache (Cached)
```

**Check the network I/O**

```bash
 sar -n DEV 1 #Report Network device statistics
 sar -n TCP,ETCP,DEV 1 #Report the incoming and outgoing tcp connections along with network device stats.
```

**Check the network performance**

```bash
netstat -s
netstat -tnlp
netstat -r #prints the route table
#Get to know the things happening at the tcp level.
#Check with applications are using which ports
```

**Check the network interface performance**

```bash
nicstat 1
#Get the network interface statistics
#check the network throughput (read, writes) and interface %util
```

**Check the process stats**

```bash
pidstat -t 1 #%user, %system
pidstat -d 1 #Disk I/O
#Very useful process stats, e.g. by thread or disk I/O
```

**Trace System Calls**

```bash
strace -p <pid>
strace -tp ‘pgrep <process>’ 2>&1 | head -100
#system call tracer - basically records all the system calls going on.
```

**Trace Network Packets**

```bash
tcpdump 
```

**Trace Kernel Message**

```bash
dmesg | tail
#Check if there are any errors
```

References: I have compiled this list through Brendan Greggs's talk and slide deck. He is a legend in this space. Find him [here](https://www.brendangregg.com/).