I'm noticing that my servers are consistently running above 90% CPU/RAM usage. In these situations, what do you usually check? Should I focus on logging, process analysis, or third-party tools? As a beginner, where should I start and what metrics should I pay attention to?
How do I find out what's consuming too many resources on my cloud server?
👁️ 5 views💬 4 replies❤️ 0 likes
4 Replies
Had the exact same issue on my old homelab setup—ended up comparing it to my iPhone getting sluggish when I’ve got 20 tabs open with dark mode + background refresh lol. First thing I did was check Activity Monitor (Windows Task Manager equivalent on macOS) to catch any rogue processes hogging CPU—turned out a misconfigured cron job was spawning Python scripts like some kind of digital gremlin. Switched to `htop` after that since it’s more visual and easier to sort by resource usage on Linux.
After that, I ran `dmesg | grep -i error` to see if the kernel was throwing a fit (like when iOS restarts randomly and you lose your Spotify queue). For long-term tracking, stuck with Netdata—way more lightweight than Grafana for a beginner yet gives you real-time heatmaps of disk, CPU, and RAM just like how Apple’s Activity Monitor shows app impact but better organized. If you’re on AWS/GCP, their built-in CloudWatch or Cloud Monitoring tools are the cloud equivalent of checking "Battery Health" in settings—shows anomalies before you even notice slowdowns.
Well, buddy, I also initially faced a similar situation where things were constantly running in the background of my smartphone. Let me break it down step by step for you, I think you should start here:
First, monitor the running processes in the bash terminal using the `top` command. If you have `htop` installed (it's a more advanced and user-friendly version), that's even better. Look for processes consuming high CPU/RAM. For example, in my case, `mysqld` or `node` processes were always spiking, and I found the issue there. Then, check disk I/O with `iotop` because sometimes processes can hog disk resources, causing the system to slow down.
When it comes to logging, there are system-related log files under `/var/log/`. You can read system logs with the `journalctl -xe` command, especially searching for words like "failed" or "error". Additionally, use the `lsof -i` command to see which ports are open and which processes are using them. As for third-party tools, `glances` gives you everything in one command—I think it's ideal for beginners.
Well, at first I was stuck just like you, then I did something simple, like "I need to see how it's done." First thing, you'll install htop, it's as simple as a command like `sudo apt install htop`, you upload it to your server and run it, and you can see all the processes live. It's immediately clear which process is consuming the most CPU, then you need to check its logs one by one.
I think sometimes unnecessary or forgotten services in the system are also consuming resources, for example, let's say you used to use Docker containers but forgot about them, you can check them with the `docker stats` command. Also, don't skip log files, especially check `/var/log/syslog` or the application's own log folder, maybe there's a loop or something constantly writing the same error. Lastly, there are third-party tools like Netdata, it's easy to install and you can monitor the entire system with its web interface, I've found it quite useful too.
Oh man, I was in that exact situation last month, bro. First, I checked which processes were eating up resources using the `top` command, then `htop` made it way easier to see. I looked at the logs and realized multiple Node.js apps were fighting over the same resources—honestly, I got a little freaked out. After optimizing the resources, everything went back to normal.