How to Keep Your Oracle Cloud Instance Alive Using stress-ng

The Problem

Oracle Cloud monitors resource usage on Always Free instances. If your server sits idle for too long — meaning CPU usage stays near zero — Oracle may flag it for reclamation or shut it down entirely.

The fix is straightforward: keep the instance visibly active by running a lightweight, continuous load in the background using stress-ng.

Requirements

• An Ubuntu instance on Oracle Cloud (Always Free or paid)
• sudo access

Step 1: Install stress-ng

stress-ng is a tool designed to stress-test system components by generating synthetic load on the CPU, memory, and other subsystems.

sudo apt update && sudo apt install stress-ng -y

Step 2: Create a systemd Service

Rather than running the command manually every time, we’ll create a systemd service that starts automatically on boot and stays running indefinitely.

sudo nano /etc/systemd/system/stress-cpu.service

Paste the following:

[Unit]
Description=CPU and RAM Keepalive Service for Oracle
After=network.target

[Service]
Type=simple
ExecStart=/usr/bin/stress-ng --cpu 2 --cpu-load 15 --vm 1 --vm-bytes 20%
Restart=always
RestartSec=3

[Install]
WantedBy=multi-user.target

Option Breakdown

Option	Description
--cpu 2	Spawn two CPU worker processes
--cpu-load 15	Target 15% CPU load per worker
--vm 1	Spawn one memory worker process
--vm-bytes 20%	Allocate 20% of total RAM
Restart=always	Automatically restart if the service crashes
RestartSec=3	Wait 3 seconds before restarting

15% CPU load is enough to register as meaningful activity without affecting the performance of anything else running on the server.

Step 3: Enable and Start the Service

Reload systemd to pick up the new service file, then enable and start it:

sudo systemctl daemon-reload
sudo systemctl enable stress-cpu
sudo systemctl start stress-cpu

• daemon-reload — tells systemd to re-read its configuration and recognize the new service.
• enable — marks the service to start automatically on every boot.
• start — starts it immediately without requiring a reboot.

Step 4: Verify It’s Running

sudo systemctl status stress-cpu

You should see output similar to this:

● stress-cpu.service - CPU and RAM Keepalive Service for Oracle
     Loaded: loaded (/etc/systemd/system/stress-cpu.service; enabled)
     Active: active (running) since ...

If the status shows failed instead of active (running), verify the path to stress-ng by running which stress-ng and make sure it matches the path in ExecStart.

Monitoring the Load with htop

Once the service is running, open htop to confirm the load is being applied:

htop

You should notice:

• CPU bars showing steady activity around 15%
• A slight increase in RAM usage from the --vm worker

This confirms the service is working as intended.

Conclusion

With a single service file and a few commands, your Oracle Cloud instance will maintain a consistent level of activity that keeps it off Oracle’s idle radar. The load is light enough to have no real impact on your workloads, and persistent enough to register as genuine resource usage — exactly what’s needed to protect an Always Free instance.