How do you manage and monitor processes and resources in CLI?

CLI tools manage local and cloud resources. Basic commands (top, ps, kill) for system processes, AWS CLI for cloud services, and kubectl for Kubernetes environments. These provide comprehensive management across systems, cloud, and containerized applications.

CLI process management is crucial for system optimization. Utilize powerful tools like top, htop, ps, and kill to monitor CPU, memory, and disk usage. Identify resource-hungry processes, prioritize tasks, and efficiently allocate resources. Real-time insights are essential. Leverage vmstat, iostat, and sar for in-depth system performance analysis. Detect bottlenecks, optimize workloads, and ensure optimal system health. Proactive management is key. Implement regular monitoring, scripting, and automation. Create custom scripts to automate routine tasks, set up alerts for critical thresholds, and prevent performance degradation.

1. ps Command: Use ps to view current processes. Customize output with options like ps -u username for a specific user, ps -e for all processes, and ps -f for detailed format. 2. top Command: top provides a dynamic, real-time view of processes, sorted by CPU or memory usage. Navigate using arrow keys, refresh with the space bar, get help with h, and quit with q. 3. Filtering and Customizing: Combine ps and top with other CLI tools and scripts to filter, sort, and manage process information, enhancing your ability to monitor system performance.

While many aspects of the CLI are harmonized my POSIX, there are examples of critical concepts that differ significantly between Unix platforms: 'killall' vs. 'killall' - In BSD and Linux, 'killall' will target specific processes by name, making it easy to shut down problematic applications. - In contrast on System V UNIX, the 'killall' command literally kills all available processes. Leaving only the kernel and the current process. It will bring down multiuser mode or the entire system. (On Linux it is called 'killall5') /proc filesystem: - On Linux the directory /proc contains a virtual file system that provides system information. Tools like 'top' and 'ps' rely on /proc. - Other UNIX systems like Macos completely lack this feature.

I used often these commands when working for bug fixing as a developer in an Oracle Sun Solaris project. There were times when debugging in investigating a defect led to blocking the system and you had to do a kill PID to forcefully terminate the process, or a kill -9 to kill all processes and start again with a clean environment. To find the PID I used ps command. Windows has Task Manager as UI, where you can terminate the processes by clicking on them, you can do the same using CLI, without the UI, so it is good to learn bash and PowerShell to use them; they work faster than using the UI, this is a useful tip when your system is overloaded or to avoid overloading.

You need free to check how you stand with the memory usage especially when your application is a big consumer, or when you want to optimize resources. Another resource is storage space, witch nowadays is big enough to cover your needs in most of the common situations. Only for enterprise distributed systems or on issues when migrating systems, for example, you have to be carefully with the storage resources. In cloud computing systems, memory and storage must be monitored also, because cost parameter must be optimized as you pay-as-you-go.

When working remotely I used ipconfig, ping and also traceroute to seek network information, test the connectivity and determine latency or the bottlenecks often due to lack of access rights on servers in the path to my target workstation.

It is noteworthy, that not only the historic 'top' implementation, but also the less-antique variant 'htop' are becoming superseded by more. Popular CLI-based resource monitors include: - atop: Advanced system and process monitor for linux systems, using process events -glances: Comprehensive alternative to top/htop - btop: State of the art resource monitor, C++ version and continuation of bashtop and bpytop -nettop: Network monitor, flow-centered similar to tcpdump The latest Macos platforms no longer provide the ?'dtrace' facility required for the mentioned 'iotop'. Modern alternatives are available via the homebrew package management: - mactop: Top monitor for Apple Silicon written in Golang, including GPU and power metrics

nice and renice: the nice command to change a process's priority, and the renice command to change the priority of multiple processes.

ping and vmstat: The ping command tests the reachability of a host on an IP network by sending ICMP echo request packets and waiting for a reply. It measures the round-trip time and reports any packet loss. For example, to ping www.example.com, you would enter ping www.example.com. The vmstat command reports information about system resources such as processes, memory, and CPU activity. The first report shows averages since the last reboot and can help identify system bottlenecks. For example, to display 3 reports every 2 seconds, you would enter vmstat 2 3.

The diverse commands: netstat, ss, lsof, etc. that can help to identify what process run in a determined network port (TCP/UDP).

traceroute is useful when working remotely across the countries. Command tracks the path that data packets take from your laptop to the server you want to login to work. It helps identifying how long data travels and where it stops if bottlenecks. I could see where network configurations or firewalls block my access to certain servers, and I asked for permissions to pass, then retried and I could see how it took me seamless to my targeted machine.