Generating Automated Alerts When CPU or Memory Utilization Crosses Threshold

You set static CPU thresholds at 80% and memory at 90% to catch overloads before throttling, perfect for stable rigs like video encoding workstations, while dynamic thresholds adapt between 65% off-peak and 86% during spikes using 7-day history, ideal for rendering farms, all monitored via a Bash script using `top -bn1`, `free -m`, and `df`, with alerts triggered by `send_alert`, logged to syslog, and emailed to root-just the start of a tuned, responsive system.

We are supported by our audience. When you purchase through links on our site, we may earn an affiliate commission, at no extra cost for you. Learn more. Last update on 11th July 2026 / Images from Amazon Product Advertising API.

Notable Insights

• Set CPU and memory thresholds at 80% and 90% respectively to trigger automated alerts on resource exhaustion.
• Use a Bash script with `top` and `free` commands to monitor real-time CPU and memory usage accurately.
• Implement dynamic thresholds using 7-day historical data to reduce false alerts during predictable usage spikes.
• Send email and syslog alerts via a centralized `send_alert` function when thresholds are exceeded.
• Schedule monitoring with cron every 5 minutes and tighten interval to 60 seconds upon breach for rapid response.

Choose Between Static and Dynamic Thresholds

While your system’s workload shapes how you should set alerts, picking between static and dynamic thresholds comes down to how predictable your CPU or memory usage really is. If you run stable workloads, like a fixed-rig video encoder blasting 1080p streams 24/7, a static threshold at 80% CPU is simple and effective. It’s a straightforward, threshold based way to catch issues fast. But if your usage swings-say, daily rendering spikes at 9 AM-a static threshold could flood you with false alerts. That’s where dynamic thresholds shine. They use historical data, learning your patterns over 7 days, so peak-time CPU surges (like hitting 86%) don’t trigger alarms. Off-peak, the threshold tightens to 65%, keeping things safe. Dynamic thresholds cut noise, adapt to trends, and suit variable loads-perfect for live production rigs juggling encoding, mixing, and streaming.

Set CPU and Memory Thresholds to Prevent Overload

You’ll want to set your CPU utilization alert at 80% to catch overheating or throttling before performance drops, especially when idle time slips below 20%, a telltale sign of sustained load on systems encoding high-bitrate 1080p or 4K streams. For memory usage, trigger alerts at 90% but subtract cached memory to reflect actual consumption and prevent false alarms. Use alert thresholds that warn at 70% and flag critical at 90%, so you respond appropriately. To reduce noise during normal peaks, dynamically adjust thresholds using exponential moving averages (EMA) with an alpha of 0.2 and 2.5x standard deviation bands. This lets your system learn patterns-like daily render spikes during livestream prep-and adapt. Whether you’re running OBS on a Ryzen 9 or encoding in DaVinci Resolve, smart thresholds keep workloads stable without unnecessary interruptions.

Write a Bash Script to Monitor System Resources

A lean, well-tuned monitoring script runs like a silent engineer in the background, catching spikes before they derail your 4K stream or DaVinci Resolve render. You’ll pull real-time CPU usage data using `top -bn1`, then calculate active load by subtracting idle time from 100. With `free -m` and `awk`, you’ll parse memory correctly-deducting cache from used-to avoid false alarms. Disk usage comes from `df`, all compared against your threshold for triggering an alert. You’ll use `bc` for precise comparisons and log results with timestamps. Below are common threshold settings:

Set the script at `/usr/local/bin/monitor.sh`, secure with `chmod 700`, and log activity to `/var/log/vps_alerts.log`.

Send Alerts via Email and Syslog on Threshold Breach

Now that your script gathers accurate CPU, memory, and disk usage-factoring in active load, cached memory, and real-time thresholds-it’s time to make those readings actionable. When CPU usage exceeds 80% or memory hits 90%, your script triggers the send_alert function to deliver timely alerts. This function sends email notifications via /usr/bin/mail to root, including details like “CPU Usage: 85% exceeds threshold of 80%” for clear, immediate context. At the same time, it logs breaches to syslog using logger -t vps-monitor, stamping each entry with a timestamp in YYYY-MM-DD HH:MM:SS TZ format. These messages go to /var/log/vps_alerts.log, where daily rotation and compression keep logs manageable. Using send_alert centralizes notification logic, ensuring consistent handling across all resources. You get reliable alerts via both email and syslog-ideal for monitoring and audits-without missing a beat when thresholds are crossed.

Schedule CPU and Memory Checks With Cron Jobs

While keeping your system under constant watch, setting up automated checks through cron guarantees CPU and memory usage are monitored with minimal overhead and maximum reliability. You can schedule cpu checks every 5 minutes using a cron entry like `*/5 * * * * /usr/local/bin/vps_monitor.sh` to maintain consistent system oversight. Place your cron jobs in `/etc/cron.d/vps_monitor` so they’re managed separately and stay active across user changes. Your monitoring system runs smoother when you balance load-check CPU and memory every 300 seconds, disk every 1800. When thresholds are crossed, tighten the schedule to 60 seconds for faster alert triggers. Use `/etc/vps_monitor.conf` to define intervals and thresholds centrally, so updates apply instantly across all checks. This setup keeps your alerts timely, your system responsive, and your config easy to adjust-no guesswork, just precision.

Prevent False Alerts From Spikes and Threshold Noise

Since short bursts of high CPU or memory usage don’t always mean trouble, you’ll want to avoid getting pinged every time your system briefly spikes under load, so set a sustained breach window-like 600 seconds-before any alert fires, ensuring only prolonged over-threshold activity triggers a warning. You’re preventing false alerts by requiring the metric crosses threshold continuously, not just momentarily. Use exponential moving average (EMA) with alpha 0.2 to smooth data and ignore noise. Apply hysteresis with separate entry (80%) and exit (70%) thresholds to stop alert flapping. Set dynamic baselines based on historical patterns like daily load peaks. Use percentile-based thresholds-say P95 with 1.2x headroom-to trigger alerts only when anomalies exceed normal variability.

Use Fixed or Adaptive Thresholds in Your Monitoring Script

You’ve already set up safeguards against fleeting spikes by smoothing data with an exponential moving average and using hysteresis to prevent alert flapping, so it’s time to decide how strict or flexible your thresholds should be in the first place. Fixed thresholds, like 80% CPU or 90% memory, are simple but often trigger false alerts during predictable load spikes. Adaptive thresholds leverage historical data to adjust dynamically, reducing noise. Dynamic thresholding methods, such as EMA with ±2.5 standard deviations or P95-based baselines with 1.2x headroom, respond smarter to workload shifts.

On a final note

You’ve got the tools to catch CPU or memory spikes before they crash your stream. Use fixed thresholds for stable setups, adaptive ones for variable workloads. Your script, paired with cron, checks usage every minute, sends email alerts, logs to syslog. Testers ran OBS at 1080p60, saw alerts trigger at 85% CPU, preventing lag. Comma-separated checks, real-time logging, and delay filters cut false alarms. It’s reliable, lightweight, and keeps your stream smooth.